Replicas
Desktop app for Mac, Windows and Linux. Agents run on your own server or in the vendor's cloud.
Product websitereplicas.dev Compare with others Missing or wrong data?
- Desktop app
- Mac, Windows and Linux
- Browser
- Yes
- Phone
- IPhone and phone browser
- Command line
- Yes
- Agents
- Claude Code, Codex, Cursor Agent, OpenCode and Pi
- Your subscription
- Works with your own plan
- Agents run
- On your own server and in the vendor's cloud
- Laptop closed
- Keeps going
- Isolation
- VM
- Parallel agents
- Yes
- Scheduled runs
- Yes
- Coordinator agent
- Yes
- Built-in issues
- Yes
- Pull requests
- Yes
- Review screen
- Yes
- Team
- Shared sessions and roles
- Self-hosting
- Yes
- Licence
- Proprietary
- Price from
- $41–100/month
- How it charges
- Subscription, usage-based, per machine or compute and bring your own key
- Paid plans from
- $41–100/month
Receipts
- Checked
- · oldest answer 2026-11-07
- Findings
- 48 published · 32 cross-checked by a second agent · read them all
- Sources
- 29 public pages on 2 sites: docs.replicas.dev, replicas.dev
- Screenshot
- replicas.dev, retrieved · first party marketing · shows appearance, not capability
- Data
- /products/replicas.json · .md
- Method
- Same rules for all 103 researched ADEs. How we check · Report wrong data
What it does
Where you use it
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
Source and scope for Where you use itAgents it runs
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
Source and scope for Agents it runsAccounts and subscriptions
Own coding-agent subscription or API credentials are required in addition to Replicas account.
Source and scope for Accounts and subscriptionsWhere agents run
Managed cloud VMs run independently of clients; Enterprise can run the platform in own infrastructure.
Source and scope for Where agents runHow work is kept apart
Separate Linux VM per workspace is documented; chats within one VM share its working files.
Source and scope for How work is kept apartParallel work
Concurrent chats/workspaces and separate-VM ships run parallel work.
Source and scope for Parallel workScheduled runs
API/Automations run saved tasks without an open client.
Source and scope for Scheduled runsGit and pull requests
Branches, PR/MR creation, review, checks, conflicts and merge operate inside product.
Source and scope for Git and pull requestsLicence and source
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
Source and scope for Licence and sourcePrice
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
Source and scope for PriceTyped filters
Every value links to its receipt below. Read plan and surface restrictions before using a filter to decide.
| Filter | Value |
|---|---|
| Desktop OS: macOS | Yes |
| Desktop OS: Windows | Yes |
| Desktop OS: Linux | Yes |
| Web client | Yes |
| Phone access: Native iOS | Yes |
| Phone access: Native Android | No |
| Phone access: Phone browser | Yes |
| CLI | Yes cross-checked |
| Coding agents: Claude Code | Yes |
| Coding agents: Codex | Yes |
| Coding agents: Cursor Agent | Yes |
| Coding agents: Gemini CLI | No |
| Coding agents: OpenCode | Yes |
| Coding agents: Pi | Yes |
| Coding agents: Grok | No |
| Coding agents: Aider | No |
| Coding agents: Antigravity | No |
| Use my subscription | Yes cross-checked |
| Where agents run: On your computer | No |
| Where agents run: On your own server | Yes cross-checked |
| Where agents run: In the vendor's cloud | Yes cross-checked |
| Workspace isolation: Git worktree | Unknown |
| Workspace isolation: Container | Unknown |
| Workspace isolation: VM or microVM | Yes cross-checked |
| Parallel runs | Yes cross-checked |
| Coordinator | Yes cross-checked |
| Native issue tracking | Yes |
| Unattended triggers | Yes cross-checked |
| Shared prompting | Yes cross-checked |
| Team roles | Yes cross-checked |
| SSO | Unknown |
| Pull request workflow | Yes cross-checked |
| Review interface | Yes cross-checked |
| Product license: Apache-2.0 | No. Proprietary |
| Product license: MIT | No. Proprietary |
| Product license: GPL-3.0 | No. Proprietary |
| Product license: GPL-3.0-or-later | No. Proprietary |
| Product license: AGPL-3.0 | No. Proprietary |
| Product license: AGPL-3.0-or-later | No. Proprietary |
| Product license: Elastic License 2.0 | No. Proprietary |
| Product license: Custom terms | No. Proprietary |
| Product license: Proprietary | Yes. Proprietary |
| Self-hosting | Yes cross-checked |
| Telemetry opt-out | Unknown |
| Free access | No cross-checked |
| Starting price: Free | No. $41–100/month cross-checked |
| Starting price: Up to $20/month | No. $41–100/month cross-checked |
| Starting price: $21–40/month | No. $41–100/month cross-checked |
| Starting price: $41–100/month | Yes. $41–100/month cross-checked |
| Starting price: Over $100/month | No. $41–100/month cross-checked |
| Starting price: Contact for pricing | No. $41–100/month cross-checked |
| Pricing model: Free, no paid plan | No |
| Pricing model: Free plan | No cross-checked |
| Pricing model: Subscription | Yes cross-checked |
| Pricing model: Per seat | No cross-checked |
| Pricing model: Usage-based | Yes cross-checked |
| Pricing model: Per machine or compute | Yes cross-checked |
| Pricing model: Bring your own key | Yes cross-checked |
| Pricing model: One-time purchase | No |
| Paid plans from: No paid plan | No. $41–100/month |
| Paid plans from: $10 or less/month | No. $41–100/month |
| Paid plans from: $11–20/month | No. $41–100/month |
| Paid plans from: $21–40/month | No. $41–100/month |
| Paid plans from: $41–100/month | Yes. $41–100/month |
| Paid plans from: Over $100/month | No. $41–100/month |
| Paid plans from: Contact sales | No. $41–100/month |
| Lifecycle status: Active | Unknown |
| Lifecycle status: Sunsetting | Unknown |
| Keeps going with laptop closed | Yes cross-checked |
| Open-source license | No. Proprietary |
| Any desktop app | Yes |
| Paid plan up to $20/month | No. $41–100/month |
| Paid plan up to $40/month | No. $41–100/month |
| Any phone client | Yes |
Filter receipts
Desktop OS: macOS: Yes
Desktop OS: macOS
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Desktop os: macOS12+, Windows10+, Linux x64/Arm.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Browser: Hosted app browser interface.
- Cli: replicas CLI.
- Task actions: iOS starts tasks/queues followups/reviews changes and Canvas.
- Notifications: macOS native turn notifications.
- Cross device continuity: Cloud task persists when iOS closes; shared account workspaces.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
Desktop OS: Windows: Yes
Desktop OS: Windows
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Desktop os: macOS12+, Windows10+, Linux x64/Arm.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Browser: Hosted app browser interface.
- Cli: replicas CLI.
- Task actions: iOS starts tasks/queues followups/reviews changes and Canvas.
- Notifications: macOS native turn notifications.
- Cross device continuity: Cloud task persists when iOS closes; shared account workspaces.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
Desktop OS: Linux: Yes
Desktop OS: Linux
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Desktop os: macOS12+, Windows10+, Linux x64/Arm.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Browser: Hosted app browser interface.
- Cli: replicas CLI.
- Task actions: iOS starts tasks/queues followups/reviews changes and Canvas.
- Notifications: macOS native turn notifications.
- Cross device continuity: Cloud task persists when iOS closes; shared account workspaces.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
Web client: Yes
Web client
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Desktop os: macOS12+, Windows10+, Linux x64/Arm.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Browser: Hosted app browser interface.
- Cli: replicas CLI.
- Task actions: iOS starts tasks/queues followups/reviews changes and Canvas.
- Notifications: macOS native turn notifications.
- Cross device continuity: Cloud task persists when iOS closes; shared account workspaces.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
Phone access: Native iOS: Yes
Phone access: Native iOS
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Desktop os: macOS12+, Windows10+, Linux x64/Arm.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Browser: Hosted app browser interface.
- Cli: replicas CLI.
- Task actions: iOS starts tasks/queues followups/reviews changes and Canvas.
- Notifications: macOS native turn notifications.
- Cross device continuity: Cloud task persists when iOS closes; shared account workspaces.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
Phone access: Native Android: No
Phone access: Native Android
Typed value: No.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Replicas: Native mobile distribution and requirements
Finding: The complete Mobile app offering is for iPhone/iPad through the App Store and requires iOS 26+. The page offers no native Android client.
Phone access: Phone browser: Yes
Phone access: Phone browser
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Mobile browser: The current changelog explicitly fixes phone-browser text-field zoom.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Replicas: Changelog phone-browser text-field zoom fix; Mobile app distribution
Finding: The current changelog explicitly fixes phone browsers zooming into text fields, establishing an intentional phone-browser client. The complete native Mobile app distribution is for iPhone/iPad on iOS 26+ through the App Store; no Android client is offered.
CLI: Yes cross-checked
CLI
Typed value: Yes.
CLI, versioned REST API and MCP control workspaces; sanctioned unattended paths are API and Automations.
- Cli: Product-owned CLI.
- Api: Bearer-token REST endpoints.
- Mcp server: Remote MCP OAuth or bearer key; local wrapper forwards.
- Webhooks: Ready/turn-completed callbacks.
- Operations: Create/message/interrupt/read/share/workspace lifecycle, environments and automations.
- Authentication: Personal or org API keys; OAuth MCP approvals revocable.
- Stability: API version header documented; terms restrict scripted interactive surfaces.
- Not established by this source pass: sdk.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
-
Finding: Hosted MCP exposes workspace/chat/environment/repository/automation/learning controls with same API permissions; OAuth creates revocable personal key.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Coding agents: Claude Code: Yes
Coding agents: Claude Code
Typed value: Yes.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Session integration: Each chat retains agent/model/effort across wakeups.
- Mixed harness operation: Mothership can choose different harness/model/effort per ship; Cursor cannot orchestrate.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Coding agents: Codex: Yes
Coding agents: Codex
Typed value: Yes.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Session integration: Each chat retains agent/model/effort across wakeups.
- Mixed harness operation: Mothership can choose different harness/model/effort per ship; Cursor cannot orchestrate.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Coding agents: Cursor Agent: Yes
Coding agents: Cursor Agent
Typed value: Yes.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Session integration: Each chat retains agent/model/effort across wakeups.
- Mixed harness operation: Mothership can choose different harness/model/effort per ship; Cursor cannot orchestrate.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Coding agents: Gemini CLI: No
Coding agents: Gemini CLI
Typed value: No.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Harnesses: The current agent/default picker is the fixed six-agent inventory; it omits Gemini CLI, Grok, Aider and Antigravity.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Replicas: Welcome and organization Default Coding Agent inventory
Finding: The current first-party documentation names the full coding-agent picker: Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi. The defaults page applies that same picker to app, API, Slack, Linear, GitHub, GitLab and automations; Gemini CLI, Grok, Aider and Antigravity are omitted. Plugins extend skills/MCP for existing agents, not the harness picker.
Coding agents: OpenCode: Yes
Coding agents: OpenCode
Typed value: Yes.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Session integration: Each chat retains agent/model/effort across wakeups.
- Mixed harness operation: Mothership can choose different harness/model/effort per ship; Cursor cannot orchestrate.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Coding agents: Pi: Yes
Coding agents: Pi
Typed value: Yes.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Session integration: Each chat retains agent/model/effort across wakeups.
- Mixed harness operation: Mothership can choose different harness/model/effort per ship; Cursor cannot orchestrate.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Coding agents: Grok: No
Coding agents: Grok
Typed value: No.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Harnesses: The current agent/default picker is the fixed six-agent inventory; it omits Gemini CLI, Grok, Aider and Antigravity.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Replicas: Welcome and organization Default Coding Agent inventory
Finding: The current first-party documentation names the full coding-agent picker: Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi. The defaults page applies that same picker to app, API, Slack, Linear, GitHub, GitLab and automations; Gemini CLI, Grok, Aider and Antigravity are omitted. Plugins extend skills/MCP for existing agents, not the harness picker.
Coding agents: Aider: No
Coding agents: Aider
Typed value: No.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Harnesses: The current agent/default picker is the fixed six-agent inventory; it omits Gemini CLI, Grok, Aider and Antigravity.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Replicas: Welcome and organization Default Coding Agent inventory
Finding: The current first-party documentation names the full coding-agent picker: Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi. The defaults page applies that same picker to app, API, Slack, Linear, GitHub, GitLab and automations; Gemini CLI, Grok, Aider and Antigravity are omitted. Plugins extend skills/MCP for existing agents, not the harness picker.
Coding agents: Antigravity: No
Coding agents: Antigravity
Typed value: No.
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Harnesses: The current agent/default picker is the fixed six-agent inventory; it omits Gemini CLI, Grok, Aider and Antigravity.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Replicas: Welcome and organization Default Coding Agent inventory
Finding: The current first-party documentation names the full coding-agent picker: Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi. The defaults page applies that same picker to app, API, Slack, Linear, GitHub, GitLab and automations; Gemini CLI, Grok, Aider and Antigravity are omitted. Plugins extend skills/MCP for existing agents, not the harness picker.
Use my subscription: Yes cross-checked
Use my subscription
Typed value: Yes.
Own coding-agent subscription or API credentials are required in addition to Replicas account.
- Agent subscriptions: Claude/Codex native subscription login via CLI.
- Api accounts: API keys, Bedrock/Foundry and supported providers.
- Local models: Codex supports Ollama/compatible Responses provider.
- Account switching: Personal/org priority and credential switch preserving history; stop chats/queue first.
- Vendor account requirement: Replicas account required.
- Provider restrictions: Mothership unavailable in Cursor; provider-specific auth applies.
- Not established by this source pass: vendor_credits.
-
Finding: Native Codex with ChatGPT subscription, own API key or compatible provider; personal/org credentials, context across sleep, compact command; capacity retry five times over about two minutes.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
Where agents run: On your computer: No
Where agents run: On your computer
Typed value: No.
Managed cloud VMs run independently of clients; Enterprise can run the platform in own infrastructure.
- Byo remote: Enterprise own-infrastructure deployment.
- Self hosted: Enterprise platform self-hosting by sales arrangement.
- Managed cloud: Replicas-hosted per-workspace LinuxVM.
- Resources: VM sizes and CPU/RAM in billing.
- Client disconnect: Cloud continues when phone/client closes.
- Other device continuation: Open same workspace from web/desktop/iOS.
- Not established by this source pass: local, hybrid, provisioning, payer, region, expiry.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Where agents run: On your own server: Yes cross-checked
Where agents run: On your own server
Typed value: Yes.
Managed cloud VMs run independently of clients; Enterprise can run the platform in own infrastructure.
- Byo remote: Enterprise own-infrastructure deployment.
- Self hosted: Enterprise platform self-hosting by sales arrangement.
- Managed cloud: Replicas-hosted per-workspace LinuxVM.
- Resources: VM sizes and CPU/RAM in billing.
- Client disconnect: Cloud continues when phone/client closes.
- Other device continuation: Open same workspace from web/desktop/iOS.
- Not established by this source pass: local, hybrid, provisioning, payer, region, expiry.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Where agents run: In the vendor's cloud: Yes cross-checked
Where agents run: In the vendor's cloud
Typed value: Yes.
Managed cloud VMs run independently of clients; Enterprise can run the platform in own infrastructure.
- Byo remote: Enterprise own-infrastructure deployment.
- Self hosted: Enterprise platform self-hosting by sales arrangement.
- Managed cloud: Replicas-hosted per-workspace LinuxVM.
- Resources: VM sizes and CPU/RAM in billing.
- Client disconnect: Cloud continues when phone/client closes.
- Other device continuation: Open same workspace from web/desktop/iOS.
- Not established by this source pass: local, hybrid, provisioning, payer, region, expiry.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Workspace isolation: VM or microVM: Yes cross-checked
Workspace isolation: VM or microVM
Typed value: Yes.
Separate Linux VM per workspace is documented; chats within one VM share its working files.
- Checkout: Repositories checked out in workspace.
- Vm: Isolated per-workspace Linux VM.
- Code separation: Separate ships get separate VM; uncommitted parent changes are not copied.
- Security boundary: Vendor asserts task VM separation; no independent security test.
- Not established by this source pass: branch, worktree, container, os_sandbox, filesystem, process, network, secrets.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Parallel runs: Yes cross-checked
Parallel runs
Typed value: Yes.
Concurrent chats/workspaces and separate-VM ships run parallel work.
- Concurrent sessions: Multiple per-workspace chats and VM workspaces.
- Concurrent tasks: Mothership ships run independent tasks.
- Quotas: Twenty unarchived ships per Mothership.
- Status visibility: Ships/Subagents running/done/failed and board.
- Not established by this source pass: observed_concurrency.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
Coordinator: Yes cross-checked
Coordinator
Typed value: Yes.
Shipped Mothership orchestration delegates separate VMs and exposes child transcripts and messages.
- Roles: Long-lived domain orchestrator and ship workers.
- Model harness effort: Inherit or override harness/model/effort; Cursor orchestration excluded.
- Messages: message_subagent and transcript tools.
- Results return: Child reports and ready/attention notifications reach parent.
- Cross harness: Choose different supported harness for ship.
- Orchestration owner: Replicas Mothership tools/board.
- Activation: Make workspace a Mothership.
- Live progress: Ships board and transcript navigation.
- Retained results: Durable notes and workspace history, subject to retention.
- Followup: Parent watches CI/review and messages ships.
- Invocation: Natural-language task delegation through Mothership.
- Not established by this source pass: authentication.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
Native issue tracking: Yes
Native issue tracking
Typed value: Yes.
Replicas Motherships own a persistent task-ticket board distinct from the derived ship/workspace and pull-request list. Tickets can be added independently, edited, ordered, archived and continued by another ship.
- Create: update_board can add tickets of its own, independently of spawning a ship.
- Edit: update_board edits ticket title, note and Linear link.
- Prioritize: update_board controls ticket ordering.
- Archive: update_board can archive tickets.
- Task session separation: A ship can continue an existing ticket through spawn_subagent ticket; the separate derived ships/PR list is below the board.
- Durable states: The product board groups tickets under Needs you, Ready to review, Working and Done.
- External tracker requirement: Linear association is optional ticket metadata; product board tickets do not require an external tracker.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: discuss, label, assign, search, filter, reopen, project_epic_views, idea_parking, explicit_dispatch, metadata.
-
Replicas: Board and Mothership tools: spawn_subagent ticket; get_board; update_board
Finding: Motherships have native board tickets with Needs you, Ready to review, Working and Done states. Each ticket can contain a summary, evidence and questions. get_board reads tickets; update_board can add independent tickets, edit title/note/Linear link, order and archive them. spawn_subagent can continue an existing ticket, while the derived ships/PR list is separately displayed below the board. This establishes persistent product-owned task records beyond the workspace/chat list.
Unattended triggers: Yes cross-checked
Unattended triggers
Typed value: Yes.
API/Automations run saved tasks without an open client.
- Schedules: Recurring schedule trigger.
- Events: Code-host/Slack/Sentry/Linear/custom webhook triggers.
- Recurring tasks: Saved prompt/environment and debounce.
- Completion conditions: Successful run limit and delete/archive/sleep/reuse lifecycle.
- Scheduler location: Replicas cloud service.
- Client independence: Triggers launch cloud work without human interaction.
- Not established by this source pass: overlaps, retries.
-
Finding: Saved prompts run on schedules or GitHub/GitLab/Slack/Sentry/Linear/custom webhooks. Fresh or reused workspace, debounce and successful-run limit; delete/archive/sleep policies; delete waits for background work.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Shared prompting: Yes cross-checked
Shared prompting
Typed value: Yes.
Shared teammates can run the same Mothership and workspaces.
- Visibility: Organization members access workspace; Shared view links.
- Discussion: Shared workspace conversation.
- Coprompting: Shared Mothership can be run together.
- Ordering: Followup messages queue FIFO when not steering.
- Attribution: Recent activity shows creator/contributors; ships use creator identity.
- Permissions: Environment and organization access policies apply.
- Not established by this source pass: comments, presence, conflicts, plan_restrictions, owner_online.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
Team roles: Yes cross-checked
Team roles
Typed value: Yes.
Organization roles and resource Read/Write apply; Enterprise advertises SCIM.
- Organization roles: Admin/Member/custom roles, additive All members.
- Project roles: Environment and automation Read/Write.
- Membership: Invitations and source-control org verification.
- Provisioning: Enterprise SCIM.
- Revocation: Removing member immediately revokes organization workspace access.
- Plan scope: SCIM is Enterprise; role docs describe current org controls.
- Not established by this source pass: task_roles, sso, unauthorized_access.
-
Finding: Admin/Member/custom roles with additive All members grants; environment/automation Read/Write and analytics Read. Removed users immediately lose organization workspaces.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Pull request workflow: Yes cross-checked
Pull request workflow
Typed value: Yes.
Branches, PR/MR creation, review, checks, conflicts and merge operate inside product.
- Branches: Each workspace has branch and repository checkout.
- Pull requests: CreatePR asks agent; GitHub/GitLab linked request UI.
- Comments: Read/post/respond or direct agent.
- Ci response: Mothership followsCI and fixes.
- Conflicts: Agent can resolve conflicts on Replicas GitHubPR.
- Product state: Header shows request state/checks and review UI.
- Not established by this source pass: worktrees, commits, rebase.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
Review interface: Yes cross-checked
Review interface
Typed value: Yes.
PR/MR tour, ordered diffs and agent chat support contextual review.
- Pending reviews: Review lists requests needing review and own requests.
- Change understanding: Tour summaries/orders files; generated/test files collapsed.
- Comments: Inline code-host comment, pending review or send to agent.
- Request changes: Request changes and direct agent fixes.
- Keyboard: ? reveals shortcuts.
- Narrow screen: Wrapped diffs and mobile Review/CreatePR bar.
- Retained context: Review workspace has checked-out PR branch.
- Not established by this source pass: evidence_navigation, shipping_guidance, effort, missed_evidence, context_switching, roles.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
Product license: Apache-2.0: No. Proprietary
Product license: Apache-2.0
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: MIT: No. Proprietary
Product license: MIT
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: GPL-3.0: No. Proprietary
Product license: GPL-3.0
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: GPL-3.0-or-later: No. Proprietary
Product license: GPL-3.0-or-later
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: AGPL-3.0: No. Proprietary
Product license: AGPL-3.0
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: AGPL-3.0-or-later: No. Proprietary
Product license: AGPL-3.0-or-later
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: Elastic License 2.0: No. Proprietary
Product license: Elastic License 2.0
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: Custom terms: No. Proprietary
Product license: Custom terms
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Product license: Proprietary: Yes. Proprietary
Product license: Proprietary
Typed value: Yes. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Self-hosting: Yes cross-checked
Self-hosting
Typed value: Yes.
Enterprise platform self-hosting and API history access are documented.
- Self hosting: Full platform in own infrastructure, Enterprise sales-scoped.
- Vendor dependencies: Selected AI provider and code-host accounts.
- Export: API reads chat/Canvas/media; complete migration export not established.
- Formats: REST JSON and Canvas/media files.
- Not established by this source pass: import, cancellation, account_independence, history_portability.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Free access: No cross-checked
Free access
Typed value: No.
Evaluation is$100 finite credit without card, not ongoing free hosted usage.
- Ongoing free: After$100 trial usage, workspace creation is blocked until subscription; no recurring free allowance.
- Trial: No time limit on unused trial credit.
- Credits trial: $100 compute trial.
- Card required: No card required.
- Usage caps: Credit exhausted stops evaluation until paid.
- Not established by this source pass: self_host_cost, paid_only, early_access, free_scenarios, paid_evaluation.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Starting price: Free: No. $41–100/month cross-checked
Starting price: Free
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Starting price: Up to $20/month: No. $41–100/month cross-checked
Starting price: Up to $20/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Starting price: $21–40/month: No. $41–100/month cross-checked
Starting price: $21–40/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Starting price: $41–100/month: Yes. $41–100/month cross-checked
Starting price: $41–100/month
Typed value: Yes. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Starting price: Over $100/month: No. $41–100/month cross-checked
Starting price: Over $100/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Starting price: Contact for pricing: No. $41–100/month cross-checked
Starting price: Contact for pricing
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Pricing model: Free, no paid plan: No
Pricing model: Free, no paid plan
Typed value: No.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Pricing model: Free plan: No cross-checked
Pricing model: Free plan
Typed value: No.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Pricing model: Subscription: Yes cross-checked
Pricing model: Subscription
Typed value: Yes.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Pricing model: Per seat: No cross-checked
Pricing model: Per seat
Typed value: No.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Pricing model: Usage-based: Yes cross-checked
Pricing model: Usage-based
Typed value: Yes.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Pricing model: Per machine or compute: Yes cross-checked
Pricing model: Per machine or compute
Typed value: Yes.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Pricing model: Bring your own key: Yes cross-checked
Pricing model: Bring your own key
Typed value: Yes.
Own coding-agent subscription or API credentials are required in addition to Replicas account.
- Agent subscriptions: Claude/Codex native subscription login via CLI.
- Api accounts: API keys, Bedrock/Foundry and supported providers.
- Local models: Codex supports Ollama/compatible Responses provider.
- Account switching: Personal/org priority and credential switch preserving history; stop chats/queue first.
- Vendor account requirement: Replicas account required.
- Provider restrictions: Mothership unavailable in Cursor; provider-specific auth applies.
- Not established by this source pass: vendor_credits.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
Pricing model: One-time purchase: No
Pricing model: One-time purchase
Typed value: No.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
Paid plans from: No paid plan: No. $41–100/month
Paid plans from: No paid plan
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Paid plans from: $10 or less/month: No. $41–100/month
Paid plans from: $10 or less/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Paid plans from: $11–20/month: No. $41–100/month
Paid plans from: $11–20/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Paid plans from: $21–40/month: No. $41–100/month
Paid plans from: $21–40/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Paid plans from: $41–100/month: Yes. $41–100/month
Paid plans from: $41–100/month
Typed value: Yes. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Paid plans from: Over $100/month: No. $41–100/month
Paid plans from: Over $100/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Paid plans from: Contact sales: No. $41–100/month
Paid plans from: Contact sales
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Keeps going with laptop closed: Yes cross-checked
Keeps going with laptop closed
Typed value: Yes.
Managed cloud VMs run independently of clients; Enterprise can run the platform in own infrastructure.
- Byo remote: Enterprise own-infrastructure deployment.
- Self hosted: Enterprise platform self-hosting by sales arrangement.
- Managed cloud: Replicas-hosted per-workspace LinuxVM.
- Resources: VM sizes and CPU/RAM in billing.
- Client disconnect: Cloud continues when phone/client closes.
- Other device continuation: Open same workspace from web/desktop/iOS.
- Not established by this source pass: local, hybrid, provisioning, payer, region, expiry.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
Open-source license: No. Proprietary
Open-source license
Typed value: No. Proprietary.
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Client license: The main platform is closed-source commercial software under the service/Enterprise terms, rather than an open-source main-application license.
- Server license: Enterprise self-hosting is supplied by custom commercial contract, not an open-source server grant.
- Source visibility: The first-party resource explicitly says Replicas is not open source and its running code cannot be read or changed.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
Any desktop app: Yes
Any desktop app
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Desktop os: macOS12+, Windows10+, Linux x64/Arm.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Browser: Hosted app browser interface.
- Cli: replicas CLI.
- Task actions: iOS starts tasks/queues followups/reviews changes and Canvas.
- Notifications: macOS native turn notifications.
- Cross device continuity: Cloud task persists when iOS closes; shared account workspaces.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
Paid plan up to $20/month: No. $41–100/month
Paid plan up to $20/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Paid plan up to $40/month: No. $41–100/month
Paid plan up to $40/month
Typed value: No. $41–100/month.
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Currency: USD.
- Interval: Monthly.
- Seats: Seats do not affect bill.
- Included usage: Subscription amount equals included usage.
- Trial: $100 usage no card/no time limit.
- Overages: Per-minute overage and periodic invoices.
- Taxes: User responsible applicable taxes.
- Model cost: Own provider subscription/API bill.
- Compute cost: VM size/minute pricing; mobile testing and GPU sandboxes billed at cost and counted toward included usage.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
Any phone client: Yes
Any phone client
Typed value: Yes.
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Desktop os: macOS12+, Windows10+, Linux x64/Arm.
- Mobile apps: The native mobile distribution is iPhone/iPad via the App Store, requiring iOS 26+.
- Browser: Hosted app browser interface.
- Cli: replicas CLI.
- Task actions: iOS starts tasks/queues followups/reviews changes and Canvas.
- Notifications: macOS native turn notifications.
- Cross device continuity: Cloud task persists when iOS closes; shared account workspaces.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
All public findings
Findings may summarize a source. They are not presented as verbatim statements.
replicas:ACC-01: Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
replicas:ACC-01
Replicas provides a web/phone-browser client, desktop apps for three OS families, CLI and native iOS access to cloud workspaces; its native mobile offering is iPhone/iPad only.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
-
Replicas: Changelog phone-browser text-field zoom fix; Mobile app distribution
Finding: The current changelog explicitly fixes phone browsers zooming into text fields, establishing an intentional phone-browser client. The complete native Mobile app distribution is for iPhone/iPad on iOS 26+ through the App Store; no Android client is offered.
-
Replicas: Native mobile distribution and requirements
Finding: The complete Mobile app offering is for iPhone/iPad through the App Store and requires iOS 26+. The page offers no native Android client.
replicas:ACC-02: Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
replicas:ACC-02
Replicas integrates the six-agent picker Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi across app and automated intake.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers, custom_adapters.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
-
Replicas: Welcome and organization Default Coding Agent inventory
Finding: The current first-party documentation names the full coding-agent picker: Claude Code, Codex, Cursor, Muse Code, OpenCode and Pi. The defaults page applies that same picker to app, API, Slack, Linear, GitHub, GitLab and automations; Gemini CLI, Grok, Aider and Antigravity are omitted. Plugins extend skills/MCP for existing agents, not the harness picker.
replicas:ACC-03: Own coding-agent subscription or API credentials are required in addition to Replicas account. cross-checked
replicas:ACC-03
Own coding-agent subscription or API credentials are required in addition to Replicas account.
- Not established by this source pass: vendor_credits.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Native Codex with ChatGPT subscription, own API key or compatible provider; personal/org credentials, context across sleep, compact command; capacity retry five times over about two minutes.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
replicas:ACC-04: Connect agent and source control, configure environment, send first task. cross-checked
replicas:ACC-04
Connect agent and source control, configure environment, send first task.
- Not established by this source pass: no_git, local_git_without_remote, byo_vps_onboarding.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
replicas:ACC-05: CLI, versioned REST API and MCP control workspaces; sanctioned unattended paths are API and Automations. cross-checked
replicas:ACC-05
CLI, versioned REST API and MCP control workspaces; sanctioned unattended paths are API and Automations.
- Not established by this source pass: sdk.
-
Finding: Product CLI creates and controls workspaces, agent authentication, shell/services/computer use.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
-
Finding: Hosted MCP exposes workspace/chat/environment/repository/automation/learning controls with same API permissions; OAuth creates revocable personal key.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
replicas:ACC-06: Environment skills/plugins/MCP tools extend selected agents. cross-checked
replicas:ACC-06
Environment skills/plugins/MCP tools extend selected agents.
- Not established by this source pass: custom_tools.
-
Finding: Hosted MCP exposes workspace/chat/environment/repository/automation/learning controls with same API permissions; OAuth creates revocable personal key.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
replicas:ACC-07: Desktop is a native window around hosted deployment; iOS includes native controls. cross-checked
replicas:ACC-07
Desktop is a native window around hosted deployment; iOS includes native controls.
- Not established by this source pass: implementation_evidence.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
replicas:PLAN-01: Prompt, integration or API starts a workspace; external tracker context can initiate work.
replicas:PLAN-01
Prompt, integration or API starts a workspace; external tracker context can initiate work.
- Not established by this source pass: local_creation, authoritative_tracker, sync, supported_edits, external_roundtrips.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Saved prompts run on schedules or GitHub/GitLab/Slack/Sentry/Linear/custom webhooks. Fresh or reused workspace, debounce and successful-run limit; delete/archive/sleep policies; delete waits for background work.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
replicas:PLAN-02: Claude/Codex plans open beside the chat in Canvas or the sidebar, with download and selection replies sent to the agent.
replicas:PLAN-02
Claude/Codex plans open beside the chat in Canvas or the sidebar, with download and selection replies sent to the agent.
- Not established by this source pass: acceptance_criteria, revisions, restart_persistence.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Claude/Codex plan sidebar with Markdown download and selection comments; Cursor plan stays in chat.
replicas:PLAN-03: Mothership/ships retain parent-child work and board state; no general issue dependency enforcement established. cross-checked
replicas:PLAN-03
Mothership/ships retain parent-child work and board state; no general issue dependency enforcement established.
- Not established by this source pass: epics, readiness, enforcement, task_to_epic, workflow_constraints.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
replicas:PLAN-04: Shared learnings, environment instructions and durable orchestrator notes preserve context. cross-checked
replicas:PLAN-04
Shared learnings, environment instructions and durable orchestrator notes preserve context.
- Not established by this source pass: project_docs, freshness.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Org skills loaded into new/resumed workspaces. Suggested persistent instructions require save; nightly recurring-instruction extraction also saves/merges learnings with audit trail. Everyone can manage; Agent Memory toggle disables.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
replicas:PLAN-05: Language-server navigation, images and MCP knowledge are available. cross-checked
replicas:PLAN-05
Language-server navigation, images and MCP knowledge are available.
- Not established by this source pass: indexing, freshness, citations, conflict_detection, review_enforcement.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
replicas:PLAN-06: Replicas Motherships own a persistent task-ticket board distinct from the derived ship/workspace and pull-request list. Tickets can be added independently, edited, ordered, archived and continued by another ship.
replicas:PLAN-06
Replicas Motherships own a persistent task-ticket board distinct from the derived ship/workspace and pull-request list. Tickets can be added independently, edited, ordered, archived and continued by another ship.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: discuss, label, assign, search, filter, reopen, project_epic_views, idea_parking, explicit_dispatch, metadata.
-
Replicas: Board and Mothership tools: spawn_subagent ticket; get_board; update_board
Finding: Motherships have native board tickets with Needs you, Ready to review, Working and Done states. Each ticket can contain a summary, evidence and questions. get_board reads tickets; update_board can add independent tickets, edit title/note/Linear link, order and archive them. spawn_subagent can continue an existing ticket, while the derived ships/PR list is separately displayed below the board. This establishes persistent product-owned task records beyond the workspace/chat list.
replicas:COORD-01: Concurrent chats/workspaces and separate-VM ships run parallel work. cross-checked
replicas:COORD-01
Concurrent chats/workspaces and separate-VM ships run parallel work.
- Not established by this source pass: observed_concurrency.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
replicas:COORD-02: Orchestrator assigns ships and exchanges followups/results. cross-checked
replicas:COORD-02
Orchestrator assigns ships and exchanges followups/results.
- Not established by this source pass: human_assignment, claims, duplicate_prevention, ownership_enforcement.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
replicas:COORD-03: Shipped Mothership orchestration delegates separate VMs and exposes child transcripts and messages. cross-checked
replicas:COORD-03
Shipped Mothership orchestration delegates separate VMs and exposes child transcripts and messages.
- Not established by this source pass: authentication.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
replicas:COORD-04: API/Automations run saved tasks without an open client. cross-checked
replicas:COORD-04
API/Automations run saved tasks without an open client.
- Not established by this source pass: overlaps, retries.
-
Finding: Saved prompts run on schedules or GitHub/GitLab/Slack/Sentry/Linear/custom webhooks. Fresh or reused workspace, debounce and successful-run limit; delete/archive/sleep policies; delete waits for background work.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
replicas:COORD-05: Steer or queue followups, stop sessions and answer questions from multiple clients.
replicas:COORD-05
Steer or queue followups, stop sessions and answer questions from multiple clients.
- Not established by this source pass: pause, escalation, visible_errors, chat_terminal_parity.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: macOS12 Intel/Arm, Windows10+ and Linux AppImage/deb x64/Arm. Native window wraps hosted deployment; mac notifications/screenshots; experimental local port forwarding.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Hosted MCP exposes workspace/chat/environment/repository/automation/learning controls with same API permissions; OAuth creates revocable personal key.
replicas:COORD-06: Ships board exposes fleet state and attention, with recent organization activity. cross-checked
replicas:COORD-06
Ships board exposes fleet state and attention, with recent organization activity.
- Not established by this source pass: fleet_fixture, dependencies, navigation_errors, fixture_mode.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
replicas:COORD-07: Single chats or Motherships support configurable workflows. cross-checked
replicas:COORD-07
Single chats or Motherships support configurable workflows.
- Not established by this source pass: growing_epic, presets, discoverability, auth_friction.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
replicas:RUN-01: Managed cloud VMs run independently of clients; Enterprise can run the platform in own infrastructure. cross-checked
replicas:RUN-01
Managed cloud VMs run independently of clients; Enterprise can run the platform in own infrastructure.
- Not established by this source pass: local, hybrid, provisioning, payer, region, expiry.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Native iOS starts cloud tasks and queues followups, reviews changes/checks/Canvas. Cloud work continues when app closes; no file browser/Code/Desktop tab. Consent before AI data sharing.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
replicas:RUN-02: Separate Linux VM per workspace is documented; chats within one VM share its working files. cross-checked
replicas:RUN-02
Separate Linux VM per workspace is documented; chats within one VM share its working files.
- Not established by this source pass: branch, worktree, container, os_sandbox, filesystem, process, network, secrets.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
replicas:RUN-03: Sleep/wake and archive/restore retain state subject to deletion and history settings.
replicas:RUN-03
Sleep/wake and archive/restore retain state subject to deletion and history settings.
- Not established by this source pass: worker_crash, service_restart, process_reconnect, state_reconstruction, duplicate_actions, manual_repair.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: History setting retains chat/Canvas asleep/archived when enabled, awake-only when disabled; repository files separate. Available trial/Team/Enterprise, not Developer.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
replicas:RUN-04: Warm/start hooks, pools, services and secrets configure development environments.
replicas:RUN-04
Warm/start hooks, pools, services and secrets configure development environments.
- Not established by this source pass: duplicate_installs, collisions, automatic_repair.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
replicas:RUN-05: Documented shell timeouts, provider retries and usage limits constrain runs. cross-checked
replicas:RUN-05
Documented shell timeouts, provider retries and usage limits constrain runs.
- Not established by this source pass: time_limits, turn_limits, token_limits, provider_failures.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Native Codex with ChatGPT subscription, own API key or compatible provider; personal/org credentials, context across sleep, compact command; capacity retry five times over about two minutes.
replicas:RUN-06: Visible VM/chat/background statuses and usage dashboards expose progress. cross-checked
replicas:RUN-06
Visible VM/chat/background statuses and usage dashboards expose progress.
- Not established by this source pass: traces, usage_export, task_stats, approval_status, reconnect_consistency, performance_method, responsiveness.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
-
Finding: Compute minutes, source, PRs/merge rate and per-seat activity; time ranges and drilldowns. Model provider costs are separate.
replicas:RUN-07: Metered VM compute and usage drilldowns are separate from own model billing.
replicas:RUN-07
Metered VM compute and usage drilldowns are separate from own model billing.
- Not established by this source pass: task_cost, session_cost, child_cost, epic_cost, harness_aggregation, billed_vs_estimated, tokens, refresh_delay, export.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
-
Finding: Compute minutes, source, PRs/merge rate and per-seat activity; time ranges and drilldowns. Model provider costs are separate.
replicas:RUN-08: CPU/memory drive automatic VM resizing; explicit broader fair scheduling not established. cross-checked
replicas:RUN-08
CPU/memory drive automatic VM resizing; explicit broader fair scheduling not established.
- Not established by this source pass: concurrency, queues, check_reuse, check_deduplication, bounded_tests, cancellation, fairness, recovery, shared_limits.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
replicas:TEAM-01: Shared teammates can run the same Mothership and workspaces. cross-checked
replicas:TEAM-01
Shared teammates can run the same Mothership and workspaces.
- Not established by this source pass: comments, presence, conflicts, plan_restrictions, owner_online.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
replicas:TEAM-02: Organization roles and resource Read/Write apply; Enterprise advertises SCIM. cross-checked
replicas:TEAM-02
Organization roles and resource Read/Write apply; Enterprise advertises SCIM.
- Not established by this source pass: task_roles, sso, unauthorized_access.
-
Finding: Admin/Member/custom roles with additive All members grants; environment/automation Read/Write and analytics Read. Removed users immediately lose organization workspaces.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
replicas:TEAM-03: Organization merge/identity policies and MCP permissions constrain actions. cross-checked
replicas:TEAM-03
Organization merge/identity policies and MCP permissions constrain actions.
- Not established by this source pass: command, tool, network, secret, approver_identity, lifetime, allow, tunnels, safer_modes.
-
Finding: Require Replicas/code-host identity and PR user attribution; block agent PR/MR merges and default protected branch pushes.
-
Finding: Hosted MCP exposes workspace/chat/environment/repository/automation/learning controls with same API permissions; OAuth creates revocable personal key.
-
Finding: Team/personal inherited environments supply variables/files/skills/MCP/plugins. Warm pools/hooks and start hooks. Read allows launching, Write editing; workspace agent gets owner access; Codex provisioned MCP auto-approves inside VM.
replicas:TEAM-04: Persistent chat and append-only configuration audit retain different histories. cross-checked
replicas:TEAM-04
Persistent chat and append-only configuration audit retain different histories.
- Not established by this source pass: approvals, decisions, cross_session_persistence, export, integrity.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: History setting retains chat/Canvas asleep/archived when enabled, awake-only when disabled; repository files separate. Available trial/Team/Enterprise, not Developer.
-
Finding: Admin-only configuration audit with actor/time/before-after JSON, secrets redacted; append-only, indefinitely retained with no delete UI.
replicas:TEAM-05: PR/MR tour, ordered diffs and agent chat support contextual review. cross-checked
replicas:TEAM-05
PR/MR tour, ordered diffs and agent chat support contextual review.
- Not established by this source pass: evidence_navigation, shipping_guidance, effort, missed_evidence, context_switching, roles.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
replicas:TEAM-06: Shared workspace/Mothership retains conversation and runtime for teammates. cross-checked
replicas:TEAM-06
Shared workspace/Mothership retains conversation and runtime for teammates.
- Not established by this source pass: decisions, artifacts, owner_disconnect, ownership, attribution, context_loss.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
replicas:SHIP-01: Canvas retains plans/media and workspace changes for review subject to retention.
replicas:SHIP-01
Canvas retains plans/media and workspace changes for review subject to retention.
- Not established by this source pass: test_outputs, reviewer_outputs, gallery, revision_links, prompt_leakage.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: History setting retains chat/Canvas asleep/archived when enabled, awake-only when disabled; repository files separate. Available trial/Team/Enterprise, not Developer.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
replicas:SHIP-02: Orchestrator follows CI/review; merge respects repository rules. cross-checked
replicas:SHIP-02
Orchestrator follows CI/review; merge respects repository rules.
- Not established by this source pass: tests, lint, types, spec_checks, required_checks, requirement_conflicts, enforced_policy.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: Require Replicas/code-host identity and PR user attribution; block agent PR/MR merges and default protected branch pushes.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
replicas:SHIP-03: Branches, PR/MR creation, review, checks, conflicts and merge operate inside product. cross-checked
replicas:SHIP-03
Branches, PR/MR creation, review, checks, conflicts and merge operate inside product.
- Not established by this source pass: worktrees, commits, rebase.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
replicas:SHIP-04: Eligible reviewers approve and merge subject to host rules and org policy.
replicas:SHIP-04
Eligible reviewers approve and merge subject to host rules and org policy.
- Not established by this source pass: release, deploy, exact_revision.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Require Replicas/code-host identity and PR user attribution; block agent PR/MR merges and default protected branch pushes.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
replicas:SHIP-05: Workspace/ships retain PR links and delivery state.
replicas:SHIP-05
Workspace/ships retain PR links and delivery state.
- Not established by this source pass: commit_links, build_links, deployment_links, rollback, revert, stale_approval_invalidation.
-
Finding: Per-workspace Linux VM; multiple harness chats; queued/steering messages; shared workspaces; file editor/diffs/inline comments, Linux desktop, terminal, Docker and Canvas. Sleep after 30 minutes inactivity, message wakes/restores; retention/deletion deadlines apply.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: History setting retains chat/Canvas asleep/archived when enabled, awake-only when disabled; repository files separate. Available trial/Team/Enterprise, not Developer.
replicas:SHIP-06: Review queue and Ships board consolidate many requests. cross-checked
replicas:SHIP-06
Review queue and Ships board consolidate many requests.
- Not established by this source pass: risk_size_filters, findings, artifacts, exact_revision_approval, dependency_queues, stale_approval_invalidation, check_deduplication.
-
Finding: Long-lived orchestrator delegates separate-VM ships, sends messages and receives results, tracks ships/PRs/Linear/blockers/decisions. Cap20 unarchived ships; ships cannot spawn ships; unavailable in Cursor chats. Shared teammates run it together.
-
Finding: PR/MR queue, guided tour ordered diffs, agent review chat, inline comments/review/agent-fix options; approve/request changes/merge subject to repository rules.
replicas:OWN-01: The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
replicas:OWN-01
The main Replicas platform is proprietary closed-source software, offered as a paid managed service or Enterprise dedicated/self-hosted deployment.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: hosted_extensions, source_revision, hosted_parity.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
-
Replicas: Where Replicas fits; closed-source limitation
Finding: The first-party self-hosting resource explicitly states that Replicas is not open source and customers cannot read or change the code being run. It describes cloud, dedicated and contract-based self-hosted Enterprise distributions of that same platform. This covers the main application rather than an SDK/release component.
replicas:OWN-02: Enterprise platform self-hosting and API history access are documented. cross-checked
replicas:OWN-02
Enterprise platform self-hosting and API history access are documented.
- Not established by this source pass: import, cancellation, account_independence, history_portability.
-
Finding: Versioned bearer-token REST API creates/messages/shares workspaces and manages automations; API-created workspaces delete when done by default unless lifecycle overridden; read never wakes.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
replicas:OWN-03: Provider processing, masked analytics, history settings and deletion are disclosed.
replicas:OWN-03
Provider processing, masked analytics, history settings and deletion are disclosed.
- Not established by this source pass: credentials, logs, training, residency, subprocessors, telemetry_defaults, telemetry_controls.
-
Finding: History setting retains chat/Canvas asleep/archived when enabled, awake-only when disabled; repository files separate. Available trial/Team/Enterprise, not Developer.
-
Finding: Selected AI providers process prompts/context. PostHog analytics/replays collect masked events; media retained beyond workspace lifetime until deletion requested; account deletion within30days.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
replicas:OWN-04: Developer starts$50/month with matching compute allowance; Team$200+; model costs separate. cross-checked
replicas:OWN-04
Developer starts$50/month with matching compute allowance; Team$200+; model costs separate.
- Not established by this source pass: seat_minimum, workspace_fees, credits.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.
replicas:OWN-06: Public signup and current weekly releases are documented.
replicas:OWN-06
Public signup and current weekly releases are documented.
- Not established by this source pass: response_commitments, status_history, deprecation, export_options.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
-
Finding: Current weekly releases include shipped Motherships for all organizations, chat tabs/harness switching and Appshots.
replicas:OWN-07: Vendor claims SOC2TypeI with TypeII in progress; report scope/period not independently verified. cross-checked
replicas:OWN-07
Vendor claims SOC2TypeI with TypeII in progress; report scope/period not independently verified.
- Not established by this source pass: iso_scope, iso_issuer, iso_validity, soc2_scope, soc2_period, report_access.
-
Finding: Enterprise offers platform self-hosted in own infrastructure or dedicated single-tenant; SCIM provisioning, audit and retention. Vendor claims SOC2 TypeI; TypeII in progress, not complete.
replicas:OWN-08: Evaluation is$100 finite credit without card, not ongoing free hosted usage. cross-checked
replicas:OWN-08
Evaluation is$100 finite credit without card, not ongoing free hosted usage.
- Not established by this source pass: self_host_cost, paid_only, early_access, free_scenarios, paid_evaluation.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
replicas:OWN-09: Public compute rates and allowance are explicit; Enterprise terms require sales.
replicas:OWN-09
Public compute rates and allowance are explicit; Enterprise terms require sales.
- Not established by this source pass: compute_markup, transparency_classification.
-
Finding: Developer$50/month includes$50 usage; Team$200+ in$100 steps, seats do not change bill. $100 trial credit no card/no time limit; exhausted trial blocked until paid. Compute rates, monthly rollover reset, daily overage invoices and hard usage limit documented.
-
Finding: Replicas account, own subscription/API credentials, GitHub/GitLab connection and environment, then send task; Claude/Codex subscription connection CLI.
-
Finding: Commercial paid hosted service; API/Automations sanctioned unattended channels. CLI interactive automation restriction differs from API documentation wording; use API/Automations for unattended claims.