Linzumi
Mac app. Agents run on your computer or on your own server.
Product websitelinzumi.com Compare with others Missing or wrong data?
- Desktop app
- Mac
- Browser
- Yes
- Phone
- Not listed
- Command line
- Yes
- Agents
- Claude Code and Codex
- Your subscription
- Works with your own plan
- Agents run
- On your computer and on your own server
- Laptop closed
- Keeps going
- Isolation
- Worktree
- Parallel agents
- Yes
- Scheduled runs
- Yes
- Coordinator agent
- Yes
- Built-in issues
- Yes
- Pull requests
- Yes
- Review screen
- Yes
- Team
- Shared sessions and SSO
- Self-hosting
- Yes
- Licence
- Not listed
- Price from
- Free
- How it charges
- Free plan, subscription and bring your own key
- Paid plans from
- $41–100/month
Receipts
- Checked
- · oldest answer 2026-11-07
- Findings
- 37 published · 30 cross-checked by a second agent · read them all
- Sources
- 10 public pages on 2 sites: linzumi.com, registry.npmjs.org
- Screenshot
- linzumi.com, retrieved · first party marketing · shows appearance, not capability
- Data
- /products/linzumi.json · .md
- Method
- Same rules for all 103 researched ADEs. How we check · Report wrong data
What it does
Where you use it
macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client.
Source and scope for Where you use itAgents it runs
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
Source and scope for Agents it runsAccounts and subscriptions
Runs use customer credentials and require a Linzumi workspace account. Commander reuses native Codex OAuth and Claude Code OAuth credentials through their own coding harnesses. The same native adapters also accept separately configured OpenAI/Anthropic provider API credentials.
Source and scope for Accounts and subscriptionsWhere agents run
Agents run on customer-controlled local or headless remote hardware.
Source and scope for Where agents runHow work is kept apart
Task Git worktrees and explicit directory permissions are documented.
Source and scope for How work is kept apartParallel work
Multiple coding agents run in parallel under team supervision.
Source and scope for Parallel workScheduled runs
Commander exposes product-owned topic wake subscriptions and scheduled checks through registered MCP handlers. Checks require the selected workspace agent runner; failure-paused subscriptions can be explicitly resumed.
Source and scope for Scheduled runsGit and pull requests
Commander can manage an existing GitHub PR as part of a coding job through native PR binding and workflow metadata. A dedicated create-PR operation is not established.
Source and scope for Git and pull requestsLicence and source
The published Commander client declares MIT and ships readable bundled implementation. The main hosted service license remains unestablished; a client package license does not settle the product-level license.
Source and scope for Licence and sourceTyped 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 cross-checked |
| Desktop OS: Windows | No |
| Desktop OS: Linux | No |
| Web client | Yes cross-checked |
| Phone access: Native iOS | Unknown |
| Phone access: Native Android | Unknown |
| Phone access: Phone browser | Unknown |
| CLI | Yes cross-checked |
| Coding agents: Claude Code | Yes |
| Coding agents: Codex | Yes |
| Coding agents: Cursor Agent | No |
| Coding agents: Gemini CLI | No |
| Coding agents: OpenCode | No |
| Coding agents: Pi | No |
| Coding agents: Grok | No |
| Coding agents: Aider | No |
| Coding agents: Antigravity | No |
| Use my subscription | Yes |
| Where agents run: On your computer | Yes cross-checked |
| Where agents run: On your own server | Yes cross-checked |
| Where agents run: In the vendor's cloud | No |
| Workspace isolation: Git worktree | Yes cross-checked |
| Workspace isolation: Container | Unknown |
| Workspace isolation: VM or microVM | Unknown |
| Parallel runs | Yes cross-checked |
| Coordinator | Yes |
| Native issue tracking | Yes |
| Unattended triggers | Yes |
| Shared prompting | Yes cross-checked |
| Team roles | Unknown |
| SSO | Yes cross-checked |
| Pull request workflow | Yes |
| Review interface | Yes cross-checked |
| Product license: Apache-2.0 | Unknown |
| Product license: MIT | Unknown |
| Product license: GPL-3.0 | Unknown |
| Product license: GPL-3.0-or-later | Unknown |
| Product license: AGPL-3.0 | Unknown |
| Product license: AGPL-3.0-or-later | Unknown |
| Product license: Elastic License 2.0 | Unknown |
| Product license: Custom terms | Unknown |
| Product license: Proprietary | Unknown |
| Self-hosting | Yes cross-checked |
| Telemetry opt-out | Unknown |
| Free access | Yes cross-checked |
| Starting price: Free | Yes. Free cross-checked |
| Starting price: Up to $20/month | No. Free cross-checked |
| Starting price: $21–40/month | No. Free cross-checked |
| Starting price: $41–100/month | No. Free cross-checked |
| Starting price: Over $100/month | No. Free cross-checked |
| Starting price: Contact for pricing | No. Free cross-checked |
| Pricing model: Free, no paid plan | No |
| Pricing model: Free plan | Yes cross-checked |
| Pricing model: Subscription | Yes cross-checked |
| Pricing model: Per seat | No cross-checked |
| Pricing model: Usage-based | No |
| Pricing model: Per machine or compute | No |
| Pricing model: Bring your own key | Yes |
| 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 | Unknown |
| Any desktop app | Yes cross-checked |
| Paid plan up to $20/month | No. $41–100/month |
| Paid plan up to $40/month | No. $41–100/month |
| Any phone client | Unknown |
Filter receipts
Desktop OS: macOS: Yes cross-checked
Desktop OS: macOS
Typed value: Yes.
macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client.
- Desktop os: macOS Apple Silicon beta download.
- Browser: Workspace chat runs in browser.
- Cli: Product-owned linzumi CLI connects and configures Commander.
- Task actions: Chat, steer, interrupt, inspect output and preview builds remotely.
- Cross device continuity: Chat client is separate from local or headless runner.
- Not established by this source pass: mobile_apps, mobile_browser, editor_extensions, notifications.
- Phone steering is advertised, but the checked first-party material does not identify the phone client.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
Desktop OS: Windows: No
Desktop OS: Windows
Typed value: No.
macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client.
- Desktop os: macOS Apple Silicon beta download.
- Browser: Workspace chat runs in browser.
- Cli: Product-owned linzumi CLI connects and configures Commander.
- Task actions: Chat, steer, interrupt, inspect output and preview builds remotely.
- Cross device continuity: Chat client is separate from local or headless runner.
- Not established by this source pass: mobile_apps, mobile_browser, editor_extensions, notifications.
- Phone steering is advertised, but the checked first-party material does not identify the phone client.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Desktop OS: Linux: No
Desktop OS: Linux
Typed value: No.
macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client.
- Desktop os: macOS Apple Silicon beta download.
- Browser: Workspace chat runs in browser.
- Cli: Product-owned linzumi CLI connects and configures Commander.
- Task actions: Chat, steer, interrupt, inspect output and preview builds remotely.
- Cross device continuity: Chat client is separate from local or headless runner.
- Not established by this source pass: mobile_apps, mobile_browser, editor_extensions, notifications.
- Phone steering is advertised, but the checked first-party material does not identify the phone client.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Web client: Yes cross-checked
Web client
Typed value: Yes.
macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client.
- Desktop os: macOS Apple Silicon beta download.
- Browser: Workspace chat runs in browser.
- Cli: Product-owned linzumi CLI connects and configures Commander.
- Task actions: Chat, steer, interrupt, inspect output and preview builds remotely.
- Cross device continuity: Chat client is separate from local or headless runner.
- Not established by this source pass: mobile_apps, mobile_browser, editor_extensions, notifications.
- Phone steering is advertised, but the checked first-party material does not identify the phone client.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
CLI: Yes cross-checked
CLI
Typed value: Yes.
Commander CLI and scoped MCP expose setup and thread context.
- Cli: linzumi connect/start/paths/mcp commands.
- Mcp server: Local linzumi MCP server preconfigured for Commander Codex.
- Operations: Read message/thread/channel and send bounded owner DM.
- Authentication: MCP uses approved local-runner authentication; headless runner receives scoped token.
- Stability: CLI release notes distinguish Commander and required server/supervisor versions.
- Not established by this source pass: api, sdk, webhooks.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
Coding agents: Claude Code: Yes
Coding agents: Claude Code
Typed value: Yes.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Coding agents: Codex: Yes
Coding agents: Codex
Typed value: Yes.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
Coding agents: Cursor Agent: No
Coding agents: Cursor Agent
Typed value: No.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Coding agents: Gemini CLI: No
Coding agents: Gemini CLI
Typed value: No.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Coding agents: OpenCode: No
Coding agents: OpenCode
Typed value: No.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Coding agents: Pi: No
Coding agents: Pi
Typed value: No.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Coding agents: Grok: No
Coding agents: Grok
Typed value: No.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Coding agents: Aider: No
Coding agents: Aider
Typed value: No.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Coding agents: Antigravity: No
Coding agents: Antigravity
Typed value: No.
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- Harnesses: Commander 1.0.340 starts both Codex and Claude Code through product-owned execution adapters.
- Custom adapters: The inspected native execution selector supports Codex and Claude Code only; generic MCP client configuration is not another coding-engine adapter.
- Session integration: Both execution adapters attach Linzumi job tools and native session/account state to the running agent.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Use my subscription: Yes
Use my subscription
Typed value: Yes.
Runs use customer credentials and require a Linzumi workspace account. Commander reuses native Codex OAuth and Claude Code OAuth credentials through their own coding harnesses. The same native adapters also accept separately configured OpenAI/Anthropic provider API credentials.
- Agent subscriptions: Commander reuses native Codex OAuth and Claude Code OAuth credentials through their own coding harnesses.
- Api accounts: The same native adapters also accept separately configured OpenAI/Anthropic provider API credentials.
- The scoped first-party source pass does not establish vendor_credits, local_models, account_switching. No hands-on scenario was run.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Where agents run: On your computer: Yes cross-checked
Where agents run: On your computer
Typed value: Yes.
Agents run on customer-controlled local or headless remote hardware.
- Local: Commander launches Codex on trusted user machine.
- Byo remote: Device-flow connect explicitly supports headless VPS/SSH machines.
- Provisioning: Connect runner with scoped device token and trusted project folders.
- Payer: User supplies hardware and credentials.
- Client disconnect: Browser control and headless runner are separate; runner terminal/supervisor must stay running.
- Other device continuation: Browser on another machine authorizes and controls remote runner.
- Not established by this source pass: self_hosted, managed_cloud, hybrid, resources, region, expiry.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
Where agents run: On your own server: Yes cross-checked
Where agents run: On your own server
Typed value: Yes.
Agents run on customer-controlled local or headless remote hardware.
- Local: Commander launches Codex on trusted user machine.
- Byo remote: Device-flow connect explicitly supports headless VPS/SSH machines.
- Provisioning: Connect runner with scoped device token and trusted project folders.
- Payer: User supplies hardware and credentials.
- Client disconnect: Browser control and headless runner are separate; runner terminal/supervisor must stay running.
- Other device continuation: Browser on another machine authorizes and controls remote runner.
- Not established by this source pass: self_hosted, managed_cloud, hybrid, resources, region, expiry.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
Where agents run: In the vendor's cloud: No
Where agents run: In the vendor's cloud
Typed value: No.
Agents run on customer-controlled local or headless remote hardware.
- Local: Commander launches Codex on trusted user machine.
- Byo remote: Device-flow connect explicitly supports headless VPS/SSH machines.
- Provisioning: Connect runner with scoped device token and trusted project folders.
- Payer: User supplies hardware and credentials.
- Client disconnect: Browser control and headless runner are separate; runner terminal/supervisor must stay running.
- Other device continuation: Browser on another machine authorizes and controls remote runner.
- Not established by this source pass: self_hosted, managed_cloud, hybrid, resources, region, expiry.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Workspace isolation: Git worktree: Yes cross-checked
Workspace isolation: Git worktree
Typed value: Yes.
Task Git worktrees and explicit directory permissions are documented.
- Worktree: Coding task checkout path under .linzumi/worktrees; uploads verify Git registration.
- Filesystem: Explicit directory ACL/trust limits folders agents can access.
- Network: Network egress requires human approval per vendor.
- Secrets: Credential reads require approval per vendor.
- Security boundary: Directory permissions and checked upload boundaries are documented; Git worktree itself is not a sandbox.
- Not established by this source pass: checkout, branch, container, vm, os_sandbox, process, code_separation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
Parallel runs: Yes cross-checked
Parallel runs
Typed value: Yes.
Multiple coding agents run in parallel under team supervision.
- Concurrent sessions: One operator/team can steer multiple agents concurrently.
- Concurrent tasks: Several coding jobs run on customer hardware.
- Status visibility: Runs are watched/logged in shared threads.
- Not established by this source pass: quotas, observed_concurrency.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Coordinator: Yes
Coordinator
Typed value: Yes.
The shipped child-job tool explicitly instructs a coding-job coordinator to plan a DAG, delegate to durable child jobs and coordinate workers through discovery and messaging.
- Messages: The coordinator discovers job reports and sends queued or urgent steering messages to workers.
- Cross harness: Child jobs can select a different supported harness/model from the coordinator.
- Orchestration owner: The product equips a coding-job agent with a coordinator role and child-job delegation tools.
- Invocation: A coordinator spawns durable child jobs under its own job using selected harness/model options.
- The scoped first-party source pass does not establish roles, model_harness_effort, results_return, activation, authentication, live_progress, retained_results, followup. No hands-on scenario was run.
-
Commander product-supplied child-job coordinator
Finding: The product supplies a coding-job coordinator tool that instructs its agent to plan a DAG, spawn durable child jobs under its own job on selected harnesses/models, discover job reports and message workers. Registered handlers dispatch through scoped Linzumi APIs; this is not merely a third-party harness subagent display. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Native issue tracking: Yes
Native issue tracking
Typed value: Yes.
The published Commander client exposes native project issue/group CRUD and dependency graphs, with durable issue state separate from coding jobs. Explicit dispatch links a project task to a coding job.
- Create: Create durable project issue/group records without launching coding work.
- Edit: Update issue titles, descriptions, notes and statuses.
- Prioritize: The task schema exposes priority and ordering.
- Assign: Project task updates assign work separately from the linked coding job.
- Explicit dispatch: Only an explicit start_coding_job with project_task_id dispatches a ready issue.
- Task session separation: Project task IDs and dependency graphs exist independently of coding jobs; dispatch creates the separate job link.
- Durable states: Issues have todo/active/blocked/review/done/canceled state and dependency-derived readiness.
- The scoped first-party source pass does not establish discuss, label, search, filter, archive, reopen, project_epic_views, idea_parking, metadata, external_tracker_requirement. No hands-on scenario was run.
-
Commander independent project issue graph
Finding: Registered live MCP handlers create durable project issue/group records with todo/active/blocked/review/done/canceled statuses, descriptions, assignees, priorities and dependency edges. Creation does not start a coding job; explicit start_coding_job binds project_task_id and transitions todo to active. This is native backlog storage separate from coding-session lifecycle. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Unattended triggers: Yes
Unattended triggers
Typed value: Yes.
Commander exposes product-owned topic wake subscriptions and scheduled checks through registered MCP handlers. Checks require the selected workspace agent runner; failure-paused subscriptions can be explicitly resumed.
- Schedules: Native scheduled topic checks run declared argv commands on a workspace agent runner.
- Events: Standing topic subscriptions wake @linzumi when entries arrive, including signed incoming webhook entries.
- Scheduler location: The configured workspace agent runner executes checks; delivery coordination uses Linzumi topic subscriptions.
- Retries: Subscriptions expose failure pause and explicit resume; this does not establish all automatic retry guarantees.
- The scoped first-party source pass does not establish recurring_tasks, completion_conditions, overlaps, client_independence. No hands-on scenario was run.
-
Commander unattended topic checks and wakes
Finding: The shipped handlers create standing subscriptions that wake @linzumi when topic entries arrive and scheduled checks that run an argv command on the workspace agent runner. Delivery supports debounce/windows, failure pause and explicit resume. Signed incoming webhook events can feed topics. This is product-owned unattended event/schedule configuration, requiring the associated runner. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Shared prompting: Yes cross-checked
Shared prompting
Typed value: Yes.
Multiple humans can prompt and steer agents in a shared thread.
- Visibility: Team watches same coding thread and output.
- Discussion: Shared channels/team threads.
- Coprompting: Humans reply, steer and interrupt the same agent thread.
- Permissions: Directory trust and operator approvals bound agent access.
- Plan restrictions: Company adds whole team; Personal is one person.
- Not established by this source pass: comments, presence, ordering, conflicts, attribution, owner_online.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
SSO: Yes cross-checked
SSO
Typed value: Yes.
Enterprise advertises SSO/SAML and SCIM; role names are not documented.
- Sso: Enterprise plan lists SSO/SAML.
- Provisioning: Enterprise lists SCIM.
- Plan scope: Enterprise procurement/contact required; no public role permission matrix.
- Not established by this source pass: organization_roles, project_roles, task_roles, membership, unauthorized_access, revocation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pull request workflow: Yes
Pull request workflow
Typed value: Yes.
Commander can manage an existing GitHub PR as part of a coding job through native PR binding and workflow metadata. A dedicated create-PR operation is not established.
- Pull requests: Native job PR binding records repository, PR, branch/head/base data and manages its workflow metadata.
- Comments: PR comments and reviews flow into the job metadata view after native binding.
- Product state: A coding job retains the primary linked GitHub PR and its workflow context.
- The scoped first-party source pass does not establish branches, worktrees, commits, ci_response, rebase, conflicts. No hands-on scenario was run.
-
Commander PR workflow metadata
Finding: The native handler links the active job to a GitHub repository/PR number and URL, stores head/base branches and SHAs, and feeds PR comments, files, commits, checks, deployments and reviews into the job workflow metadata view. This establishes management of an existing PR; it does not claim a native PR-creation command. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
Review interface: Yes cross-checked
Review interface
Typed value: Yes.
Team reviews proposed diffs and evidence inline in chat.
- Change understanding: Agent posts diff and commands run in thread.
- Evidence navigation: Thread includes files/screenshots/test output.
- Comments: Team reviews/steers inline.
- Request changes: Teammates redirect the agent mid-run.
- Not established by this source pass: pending_reviews, shipping_guidance, effort, missed_evidence, keyboard, narrow_screen, context_switching, roles, retained_context.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Self-hosting: Yes cross-checked
Self-hosting
Typed value: Yes.
Enterprise offers self-hosting; code/context export is advertised without an install recipe.
- Self hosting: Enterprise lists self-hosting/VPC/own-cloud option.
- Vendor dependencies: Personal/Company browser workspace uses hosted service plus connected Commander.
- Export: Vendor says users can leave with code and context.
- Not established by this source pass: import, formats, cancellation, account_independence, history_portability.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Free access: Yes cross-checked
Free access
Typed value: Yes.
Personal is ongoing free with no card; Company offers a first free month.
- Ongoing free: Personal free forever.
- Trial: Company first month free.
- Free scenarios: Solo coding, live preview and full replay.
- Card required: Personal requires no credit card.
- Usage caps: Personal one person/unlimited channels; own keys/hardware.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Starting price: Free: Yes. Free cross-checked
Starting price: Free
Typed value: Yes. Free.
Personal is ongoing free with no card; Company offers a first free month.
- Ongoing free: Personal free forever.
- Trial: Company first month free.
- Free scenarios: Solo coding, live preview and full replay.
- Card required: Personal requires no credit card.
- Usage caps: Personal one person/unlimited channels; own keys/hardware.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Starting price: Up to $20/month: No. Free cross-checked
Starting price: Up to $20/month
Typed value: No. Free.
Personal is ongoing free with no card; Company offers a first free month.
- Ongoing free: Personal free forever.
- Trial: Company first month free.
- Free scenarios: Solo coding, live preview and full replay.
- Card required: Personal requires no credit card.
- Usage caps: Personal one person/unlimited channels; own keys/hardware.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Starting price: $21–40/month: No. Free cross-checked
Starting price: $21–40/month
Typed value: No. Free.
Personal is ongoing free with no card; Company offers a first free month.
- Ongoing free: Personal free forever.
- Trial: Company first month free.
- Free scenarios: Solo coding, live preview and full replay.
- Card required: Personal requires no credit card.
- Usage caps: Personal one person/unlimited channels; own keys/hardware.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Starting price: $41–100/month: No. Free cross-checked
Starting price: $41–100/month
Typed value: No. Free.
Personal is ongoing free with no card; Company offers a first free month.
- Ongoing free: Personal free forever.
- Trial: Company first month free.
- Free scenarios: Solo coding, live preview and full replay.
- Card required: Personal requires no credit card.
- Usage caps: Personal one person/unlimited channels; own keys/hardware.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Starting price: Over $100/month: No. Free cross-checked
Starting price: Over $100/month
Typed value: No. Free.
Personal is ongoing free with no card; Company offers a first free month.
- Ongoing free: Personal free forever.
- Trial: Company first month free.
- Free scenarios: Solo coding, live preview and full replay.
- Card required: Personal requires no credit card.
- Usage caps: Personal one person/unlimited channels; own keys/hardware.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Starting price: Contact for pricing: No. Free cross-checked
Starting price: Contact for pricing
Typed value: No. Free.
Personal is ongoing free with no card; Company offers a first free month.
- Ongoing free: Personal free forever.
- Trial: Company first month free.
- Free scenarios: Solo coding, live preview and full replay.
- Card required: Personal requires no credit card.
- Usage caps: Personal one person/unlimited channels; own keys/hardware.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: Free, no paid plan: No
Pricing model: Free, no paid plan
Typed value: No.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: Free plan: Yes cross-checked
Pricing model: Free plan
Typed value: Yes.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: Subscription: Yes cross-checked
Pricing model: Subscription
Typed value: Yes.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: Per seat: No cross-checked
Pricing model: Per seat
Typed value: No.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: Usage-based: No
Pricing model: Usage-based
Typed value: No.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: Per machine or compute: No
Pricing model: Per machine or compute
Typed value: No.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: Bring your own key: Yes
Pricing model: Bring your own key
Typed value: Yes.
Runs use customer credentials and require a Linzumi workspace account. Commander reuses native Codex OAuth and Claude Code OAuth credentials through their own coding harnesses. The same native adapters also accept separately configured OpenAI/Anthropic provider API credentials.
- Vendor account requirement: Sign into Linzumi browser workspace to authorize Commander.
- Provider restrictions: Codex is explicitly named as Commander-started runtime; other harness execution remains unestablished.
- The scoped first-party source pass does not establish vendor_credits, local_models, account_switching. No hands-on scenario was run.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Pricing model: One-time purchase: No
Pricing model: One-time purchase
Typed value: No.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plans from: No paid plan: No. $41–100/month
Paid plans from: No paid plan
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plans from: $10 or less/month: No. $41–100/month
Paid plans from: $10 or less/month
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plans from: $11–20/month: No. $41–100/month
Paid plans from: $11–20/month
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plans from: $21–40/month: No. $41–100/month
Paid plans from: $21–40/month
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plans from: $41–100/month: Yes. $41–100/month
Paid plans from: $41–100/month
Typed value: Yes. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plans from: Over $100/month: No. $41–100/month
Paid plans from: Over $100/month
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plans from: Contact sales: No. $41–100/month
Paid plans from: Contact sales
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Keeps going with laptop closed: Yes cross-checked
Keeps going with laptop closed
Typed value: Yes.
Agents run on customer-controlled local or headless remote hardware.
- Local: Commander launches Codex on trusted user machine.
- Byo remote: Device-flow connect explicitly supports headless VPS/SSH machines.
- Provisioning: Connect runner with scoped device token and trusted project folders.
- Payer: User supplies hardware and credentials.
- Client disconnect: Browser control and headless runner are separate; runner terminal/supervisor must stay running.
- Other device continuation: Browser on another machine authorizes and controls remote runner.
- Not established by this source pass: self_hosted, managed_cloud, hybrid, resources, region, expiry.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
Any desktop app: Yes cross-checked
Any desktop app
Typed value: Yes.
macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client.
- Desktop os: macOS Apple Silicon beta download.
- Browser: Workspace chat runs in browser.
- Cli: Product-owned linzumi CLI connects and configures Commander.
- Task actions: Chat, steer, interrupt, inspect output and preview builds remotely.
- Cross device continuity: Chat client is separate from local or headless runner.
- Not established by this source pass: mobile_apps, mobile_browser, editor_extensions, notifications.
- Phone steering is advertised, but the checked first-party material does not identify the phone client.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
Paid plan up to $20/month: No. $41–100/month
Paid plan up to $20/month
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
Paid plan up to $40/month: No. $41–100/month
Paid plan up to $40/month
Typed value: No. $41–100/month.
Personal is free; Company is $100/month flat for whole team.
- Currency: USD.
- Interval: Company $100 per month.
- Seats: Company has no per-seat pricing; Personal one person.
- Included usage: Personal unlimited channels; Company unlimited agent-hours.
- Trial: Company first month free.
- Model cost: Own keys/credentials; separate agent charges not priced.
- Compute cost: Own hardware on every plan.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
All public findings
Findings may summarize a source. They are not presented as verbatim statements.
linzumi:ACC-01: macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client. cross-checked
linzumi:ACC-01
macOS Apple Silicon beta and browser chat are documented; phone steering is advertised without naming the phone client.
- Not established by this source pass: mobile_apps, mobile_browser, editor_extensions, notifications.
- Phone steering is advertised, but the checked first-party material does not identify the phone client.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:ACC-02: Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
linzumi:ACC-02
Commander 1.0.340 starts Codex and Claude Code through product-owned execution adapters. Its native selector accepts those two engines and does not expose a generic executable harness adapter. Both attach scoped Linzumi job tools and native account/session state. MCP access from a different client does not establish execution of that client as a coding harness.
- The scoped first-party source pass does not establish model_providers, mixed_harness_operation. No hands-on scenario was run.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
linzumi:ACC-03: Runs use customer credentials and require a Linzumi workspace account. Commander reuses native Codex OAuth and Claude Code OAuth credentials through their own coding harnesses. The same native adapters also accept separately configured OpenAI/Anthropic provider API credentials.
linzumi:ACC-03
Runs use customer credentials and require a Linzumi workspace account. Commander reuses native Codex OAuth and Claude Code OAuth credentials through their own coding harnesses. The same native adapters also accept separately configured OpenAI/Anthropic provider API credentials.
- The scoped first-party source pass does not establish vendor_credits, local_models, account_switching. No hands-on scenario was run.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
-
Commander native coding engines
Finding: The execution selector admits Codex and Claude Code. The Claude path imports @anthropic-ai/claude-agent-sdk, calls its query with the Claude executable and native settingSources, and accepts native OAuth credentials or provider API keys. The Codex auth path reads the native auth.json OAuth account or OPENAI_API_KEY. Other MCP client configuration formats do not launch additional engines. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
linzumi:ACC-04: Browser sign-up and trusted-folder registration precede running work; headless runner uses device flow. cross-checked
linzumi:ACC-04
Browser sign-up and trusted-folder registration precede running work; headless runner uses device flow.
- Not established by this source pass: no_git, local_git_without_remote, hosted_git, managed_cloud_setup.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:ACC-05: Commander CLI and scoped MCP expose setup and thread context. cross-checked
linzumi:ACC-05
Commander CLI and scoped MCP expose setup and thread context.
- Not established by this source pass: api, sdk, webhooks.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:ACC-06: Coding agents can access scoped Linzumi context through official MCP integration. cross-checked
linzumi:ACC-06
Coding agents can access scoped Linzumi context through official MCP integration.
- Not established by this source pass: mcp_client, plugins, instructions_import.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:PLAN-01: Shared chat starts coding work on a trusted machine. cross-checked
linzumi:PLAN-01
Shared chat starts coding work on a trusted machine.
- Not established by this source pass: local_creation, external_import, stable_ids, sync, supported_edits, external_roundtrips.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:PLAN-04: Scoped thread/channel context is available to coding agents; C3 remains roadmap. cross-checked
linzumi:PLAN-04
Scoped thread/channel context is available to coding agents; C3 remains roadmap.
- Not established by this source pass: instructions, skills, memory, provenance, correction, deletion, freshness.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:PLAN-06: The published Commander client exposes native project issue/group CRUD and dependency graphs, with durable issue state separate from coding jobs. Explicit dispatch links a project task to a coding job.
linzumi:PLAN-06
The published Commander client exposes native project issue/group CRUD and dependency graphs, with durable issue state separate from coding jobs. Explicit dispatch links a project task to a coding job.
- The scoped first-party source pass does not establish discuss, label, search, filter, archive, reopen, project_epic_views, idea_parking, metadata, external_tracker_requirement. No hands-on scenario was run.
-
Commander independent project issue graph
Finding: Registered live MCP handlers create durable project issue/group records with todo/active/blocked/review/done/canceled statuses, descriptions, assignees, priorities and dependency edges. Creation does not start a coding job; explicit start_coding_job binds project_task_id and transitions todo to active. This is native backlog storage separate from coding-session lifecycle. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
linzumi:PLAN-08: C3 context-derived specifications are announced, not a current capability. cross-checked
linzumi:PLAN-08
C3 context-derived specifications are announced, not a current capability.
- Not established by this source pass: human_edits, agent_edits, revisions, approvals, implementation_links, later_retrieval, drift_detection, review_integration.
-
Finding: C3 cited answers, live call transcription, contradiction reconciliation and context-derived verifiable specs are marked Coming soon. No release date stated.
linzumi:COORD-01: Multiple coding agents run in parallel under team supervision. cross-checked
linzumi:COORD-01
Multiple coding agents run in parallel under team supervision.
- Not established by this source pass: quotas, observed_concurrency.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:COORD-03: The shipped child-job tool explicitly instructs a coding-job coordinator to plan a DAG, delegate to durable child jobs and coordinate workers through discovery and messaging.
linzumi:COORD-03
The shipped child-job tool explicitly instructs a coding-job coordinator to plan a DAG, delegate to durable child jobs and coordinate workers through discovery and messaging.
- The scoped first-party source pass does not establish roles, model_harness_effort, results_return, activation, authentication, live_progress, retained_results, followup. No hands-on scenario was run.
-
Commander product-supplied child-job coordinator
Finding: The product supplies a coding-job coordinator tool that instructs its agent to plan a DAG, spawn durable child jobs under its own job on selected harnesses/models, discover job reports and message workers. Registered handlers dispatch through scoped Linzumi APIs; this is not merely a third-party harness subagent display. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
linzumi:COORD-04: Commander exposes product-owned topic wake subscriptions and scheduled checks through registered MCP handlers. Checks require the selected workspace agent runner; failure-paused subscriptions can be explicitly resumed.
linzumi:COORD-04
Commander exposes product-owned topic wake subscriptions and scheduled checks through registered MCP handlers. Checks require the selected workspace agent runner; failure-paused subscriptions can be explicitly resumed.
- The scoped first-party source pass does not establish recurring_tasks, completion_conditions, overlaps, client_independence. No hands-on scenario was run.
-
Commander unattended topic checks and wakes
Finding: The shipped handlers create standing subscriptions that wake @linzumi when topic entries arrive and scheduled checks that run an argv command on the workspace agent runner. Delivery supports debounce/windows, failure pause and explicit resume. Signed incoming webhook events can feed topics. This is product-owned unattended event/schedule configuration, requiring the associated runner. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
linzumi:COORD-05: Humans steer and interrupt shared runs; restoration holds remain visible. cross-checked
linzumi:COORD-05
Humans steer and interrupt shared runs; restoration holds remain visible.
- Not established by this source pass: pause, resume, notifications, escalation, chat_terminal_parity.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
linzumi:COORD-06: Threads expose a fleet of agent runs and attention states. cross-checked
linzumi:COORD-06
Threads expose a fleet of agent runs and attention states.
- Not established by this source pass: fleet_fixture, parent_child, dependencies, navigation_errors, fixture_mode.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
linzumi:COORD-07: Natural-language shared threads drive local agent work. cross-checked
linzumi:COORD-07
Natural-language shared threads drive local agent work.
- Not established by this source pass: quick_task, growing_epic, presets, agent_configuration, retained_relationships, discoverability, auth_friction.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:RUN-01: Agents run on customer-controlled local or headless remote hardware. cross-checked
linzumi:RUN-01
Agents run on customer-controlled local or headless remote hardware.
- Not established by this source pass: self_hosted, managed_cloud, hybrid, resources, region, expiry.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:RUN-02: Task Git worktrees and explicit directory permissions are documented. cross-checked
linzumi:RUN-02
Task Git worktrees and explicit directory permissions are documented.
- Not established by this source pass: checkout, branch, container, vm, os_sandbox, process, code_separation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
linzumi:RUN-03: Commander supervises local runner and recovers saved replies with receipt checks. cross-checked
linzumi:RUN-03
Commander supervises local runner and recovers saved replies with receipt checks.
- Not established by this source pass: client_disconnect, worker_crash, expiry, transcript_reload, process_reconnect, task_resume, multi_day_continuity.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
linzumi:RUN-04: Work uses existing machine environment and can forward previews. cross-checked
linzumi:RUN-04
Work uses existing machine environment and can forward previews.
- Not established by this source pass: caches, snapshots, services, secrets, duplicate_installs, collisions, automatic_repair.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:RUN-06: Run actions and recovery holds are logged and replayable. cross-checked
linzumi:RUN-06
Run actions and recovery holds are logged and replayable.
- Not established by this source pass: traces, task_stats, agent_stats, project_stats, time_stats, approval_status, reconnect_consistency, performance_method, responsiveness.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
linzumi:TEAM-01: Multiple humans can prompt and steer agents in a shared thread. cross-checked
linzumi:TEAM-01
Multiple humans can prompt and steer agents in a shared thread.
- Not established by this source pass: comments, presence, ordering, conflicts, attribution, owner_online.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:TEAM-02: Enterprise advertises SSO/SAML and SCIM; role names are not documented. cross-checked
linzumi:TEAM-02
Enterprise advertises SSO/SAML and SCIM; role names are not documented.
- Not established by this source pass: organization_roles, project_roles, task_roles, membership, unauthorized_access, revocation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:TEAM-03: User grants directories, network/credential/repo access and temporary tunnels. cross-checked
linzumi:TEAM-03
User grants directories, network/credential/repo access and temporary tunnels.
- Not established by this source pass: command, delivery, approver_identity, deny, revoke, defaults, safer_modes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:TEAM-04: Thread retains agent actions, team discussion and exportable evidence. cross-checked
linzumi:TEAM-04
Thread retains agent actions, team discussion and exportable evidence.
- Not established by this source pass: approvals, edits, decisions, cross_session_persistence, mutability, integrity.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:TEAM-05: Team reviews proposed diffs and evidence inline in chat. cross-checked
linzumi:TEAM-05
Team reviews proposed diffs and evidence inline in chat.
- Not established by this source pass: pending_reviews, shipping_guidance, effort, missed_evidence, keyboard, narrow_screen, context_switching, roles, retained_context.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:SHIP-01: Agent outputs attach to thread and may be exported. cross-checked
linzumi:SHIP-01
Agent outputs attach to thread and may be exported.
- Not established by this source pass: recordings, reviewer_outputs, gallery, revision_links, child_findings, prompt_leakage.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
linzumi:SHIP-02: Agent reports commands/tests in shared thread; enforced test gates remain unknown. cross-checked
linzumi:SHIP-02
Agent reports commands/tests in shared thread; enforced test gates remain unknown.
- Not established by this source pass: lint, types, ci, spec_checks, required_checks, failure_blocking, requirement_conflicts, configured_policy, enforced_policy.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:SHIP-03: Commander can manage an existing GitHub PR as part of a coding job through native PR binding and workflow metadata. A dedicated create-PR operation is not established.
linzumi:SHIP-03
Commander can manage an existing GitHub PR as part of a coding job through native PR binding and workflow metadata. A dedicated create-PR operation is not established.
- The scoped first-party source pass does not establish branches, worktrees, commits, ci_response, rebase, conflicts. No hands-on scenario was run.
-
Commander PR workflow metadata
Finding: The native handler links the active job to a GitHub repository/PR number and URL, stores head/base branches and SHAs, and feeds PR comments, files, commits, checks, deployments and reviews into the job workflow metadata view. This establishes management of an existing PR; it does not claim a native PR-creation command. Inspected the exact published 1.0.340 package, pinned by its npm SHA512 integrity digest rather than a Git revision.
linzumi:OWN-01: The published Commander client declares MIT and ships readable bundled implementation. The main hosted service license remains unestablished; a client package license does not settle the product-level license.
linzumi:OWN-01
The published Commander client declares MIT and ships readable bundled implementation. The main hosted service license remains unestablished; a client package license does not settle the product-level license.
- The scoped first-party source pass does not establish server_license, hosted_extensions, source_revision, hosted_parity. No hands-on scenario was run.
-
Commander package license metadata
Finding: The first-party published Commander client package declares MIT. The package does not include a license grant or source distribution for the main hosted Linzumi chat/control-plane service; its client license cannot establish the product-level hosted-core license.
linzumi:OWN-02: Enterprise offers self-hosting; code/context export is advertised without an install recipe. cross-checked
linzumi:OWN-02
Enterprise offers self-hosting; code/context export is advertised without an install recipe.
- Not established by this source pass: import, formats, cancellation, account_independence, history_portability.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:OWN-03: Vendor describes local code execution and support-channel visibility limits. cross-checked
linzumi:OWN-03
Vendor describes local code execution and support-channel visibility limits.
- Not established by this source pass: prompts, artifacts, retention, deletion, training, residency, subprocessors, telemetry_disclosure, telemetry_defaults, telemetry_controls, consent_prompt.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Finding: Official @linzumi/cli README documents trusted project folders, browser/local separation, Codex launch, headless VPS device-flow authentication, scoped MCP message/thread/channel reads and owner DM; team members can steer and interrupt the same thread.
linzumi:OWN-04: Personal is free; Company is $100/month flat for whole team. cross-checked
linzumi:OWN-04
Personal is free; Company is $100/month flat for whole team.
- Not established by this source pass: seat_minimum, workspace_fees, credits, overages, taxes.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:OWN-06: Mac app is beta; Commander package is actively published. cross-checked
linzumi:OWN-06
Mac app is beta; Commander package is actively published.
- Not established by this source pass: access_restrictions, response_commitments, status_history, deprecation, export_options.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
-
Linzumi CLI release and recovery notes
Finding: Latest package 1.0.335 published 2026-11-07. Saved replies recover only after receipt/ownership/archive checks; Ready waits for all recovery obligations. File uploads permit registered Git task worktrees and reject symlink/hidden/unrelated-path escapes.
linzumi:OWN-07: Enterprise lists SOC 2 without report type, audit period or independent verification. cross-checked
linzumi:OWN-07
Enterprise lists SOC 2 without report type, audit period or independent verification.
- Not established by this source pass: iso_scope, iso_issuer, iso_validity, soc2_type, soc2_period, report_access, verification_date.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:OWN-08: Personal is ongoing free with no card; Company offers a first free month. cross-checked
linzumi:OWN-08
Personal is ongoing free with no card; Company offers a first free month.
- Not established by this source pass: credits_trial, self_host_cost, paid_only, early_access, paid_evaluation.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.
linzumi:OWN-09: Flat plan prices are public; model/hardware cost details remain separate. cross-checked
linzumi:OWN-09
Flat plan prices are public; model/hardware cost details remain separate.
- Not established by this source pass: overages, compute_markup, model_markup, minimum_commitment.
-
Finding: macOS Apple Silicon beta and browser workspace. Multiple humans steer concurrent local agents in one thread; inline diffs, replay, previews and exportable artifacts. Personal free solo plan; Company $100/month flat; Enterprise offers SSO/SAML/SCIM and self-hosting.