Capy
Desktop app for Mac, Windows and Linux. Agents run on your computer, on your own server or in the vendor's cloud.
Product websitecapy.ai Compare with others Missing or wrong data?
- Desktop app
- Mac, Windows and Linux
- Browser
- Yes
- Phone
- Phone browser
- Command line
- Yes
- Agents
- Claude Code
- Your subscription
- Works with your own plan
- Agents run
- On your computer, on your own server and in the vendor's cloud
- Laptop closed
- Keeps going
- Isolation
- Worktree and VM
- Parallel agents
- Yes
- Scheduled runs
- Yes
- Coordinator agent
- Yes
- Built-in issues
- Yes
- Pull requests
- Yes
- Review screen
- Yes
- Team
- Shared sessions, roles and SSO
- Self-hosting
- Not listed
- Licence
- Custom terms
- Price from
- Free
- How it charges
- Free plan, subscription, usage-based, per machine or compute and bring your own key
- Paid plans from
- $11–20/month
Measured by ADEMark
ADEMark ran the same tests on 15 ADEs. Results for Capy 0.8.0, measured up to 12 November 2026 on a Mac mini M4, 16 GB, macOS 27 · Claude Opus 5.5, effort high. Data CC BY 4.0, shown unchanged.
Speed
- Cold start
- 2.01 s4th of 12median of 10 runs, 1.82 s to 2.19 s · Bare CLI 0.69 s
- New thread, ready to type
- Not measuredADEMark: it talks only to its maker's cloud, so our local replay server cannot answer it
- First word on screen
- Not measuredADEMark: it talks only to its maker's cloud, so our local replay server cannot answer it
- Typing lag, 5 agents streaming
- Not measuredADEMark: it talks only to its maker's cloud, so our local replay server cannot answer it
- Open a 2,000-message thread
- Not measuredADEMark: it talks only to its maker's cloud, so our local replay server cannot answer it
Weight
- Memory at rest, 5 threads
- 593 MB3rd of 121 run · Bare CLI 0.99 GB
- Idle CPU
- 0.56%3rd of 131 run · Bare CLI 0.37%
- Memory under load
- Not measuredADEMark: it talks only to its maker's cloud, so our local replay server cannot answer it
Tokens
- Tokens to answer “hi”
- Not measuredADEMark: it updated itself to 0.8.0 and our step that picks Opus 5.5 broke, so no prompt was sent
- Tokens for one real question
- Not measuredADEMark: it updated itself to 0.8.0 and our step that picks Opus 5.5 broke, so no prompt was sent
Build
- Team build: hidden tests passed
- Not measuredADEMark: it updated itself to 0.8.0 and our step that picks Opus 5.5 broke, so no prompt was sent
Place counts only apps with a value; lower is better except passed tests. ADEMark publishes no overall score and treats close medians as ties, so read places as an order, not a verdict. The Bare CLI is Claude Code in a terminal, run as an unranked control. Method and every test on ademark.dev · Imported data · How we use it
Receipts
- Checked
- · oldest answer 2026-11-06
- Findings
- 49 published · 44 cross-checked by a second agent · read them all
- Sources
- 51 public pages on 3 sites: docs.capy.ai, capy.ai, trust.capy.ai
- Screenshot
- docs.capy.ai, retrieved · first party docs · shows appearance, not capability
- Data
- /products/capy.json · .md
- Method
- Same rules for all 103 researched ADEs. How we check · Report wrong data
What it does
Where you use it
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
Source and scope for Where you use itAgents it runs
Capy runs its own agent plus a Claude Code mode, with a wide provider catalog; other native harnesses were not found.
Source and scope for Agents it runsWhere agents run
Cloud VMs, desktop devices, Dock machines and movable threads are documented; own-cloud is Enterprise by agreement.
Source and scope for Where agents runHow work is kept apart
Cloud threads use Firecracker microVMs; device threads use worktrees and are explicitly not sandboxed.
Source and scope for How work is kept apartParallel work
Concurrent child agents and thread-level activity aggregation are documented.
Source and scope for Parallel workScheduled runs
Stored prompts support cron, GitHub/Slack events, webhooks and standing or fresh threads.
Source and scope for Scheduled runsGit and pull requests
Branch/commit/push/PR creation, comments and CI follow-up are documented.
Source and scope for Git and pull requestsLicence and source
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
Source and scope for Licence and sourcePrice
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
Source and scope for PriceTyped filters
Every value links to its receipt below. Read plan and surface restrictions before using a filter to decide.
| Filter | Value |
|---|---|
| Desktop OS: macOS | Yes cross-checked |
| Desktop OS: Windows | Yes cross-checked |
| Desktop OS: Linux | Yes cross-checked |
| Web client | Yes cross-checked |
| Phone access: Native iOS | No cross-checked |
| Phone access: Native Android | No cross-checked |
| Phone access: Phone browser | Yes cross-checked |
| CLI | Yes cross-checked |
| Coding agents: Claude Code | Yes |
| Coding agents: Codex | Unknown |
| Coding agents: Cursor Agent | Unknown |
| Coding agents: Gemini CLI | Unknown |
| Coding agents: OpenCode | Unknown |
| Coding agents: Pi | Unknown |
| Coding agents: Grok | Unknown |
| Coding agents: Aider | Unknown |
| Coding agents: Antigravity | Unknown |
| Use my subscription | Yes cross-checked |
| 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 | Yes cross-checked |
| Workspace isolation: Git worktree | Yes cross-checked |
| Workspace isolation: Container | No |
| Workspace isolation: VM or microVM | Yes cross-checked |
| Parallel runs | Yes cross-checked |
| Coordinator | Yes cross-checked |
| Native issue tracking | Yes cross-checked |
| Unattended triggers | Yes cross-checked |
| Shared prompting | Yes cross-checked |
| Team roles | Yes cross-checked |
| SSO | Yes cross-checked |
| Pull request workflow | Yes cross-checked |
| Review interface | Yes cross-checked |
| Product license: Apache-2.0 | No. Custom terms |
| Product license: MIT | No. Custom terms |
| Product license: GPL-3.0 | No. Custom terms |
| Product license: GPL-3.0-or-later | No. Custom terms |
| Product license: AGPL-3.0 | No. Custom terms |
| Product license: AGPL-3.0-or-later | No. Custom terms |
| Product license: Elastic License 2.0 | No. Custom terms |
| Product license: Custom terms | Yes. Custom terms |
| Product license: Proprietary | No. Custom terms |
| Self-hosting | Unknown |
| 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 | Yes cross-checked |
| Pricing model: Per machine or compute | Yes cross-checked |
| Pricing model: Bring your own key | Yes cross-checked |
| Pricing model: One-time purchase | No |
| Paid plans from: No paid plan | No. $11–20/month |
| Paid plans from: $10 or less/month | No. $11–20/month |
| Paid plans from: $11–20/month | Yes. $11–20/month |
| Paid plans from: $21–40/month | No. $11–20/month |
| Paid plans from: $41–100/month | No. $11–20/month |
| Paid plans from: Over $100/month | No. $11–20/month |
| Paid plans from: Contact sales | No. $11–20/month |
| Lifecycle status: Active | Unknown |
| Lifecycle status: Sunsetting | Unknown |
| Keeps going with laptop closed | Yes cross-checked |
| Open-source license | No. Custom terms |
| Any desktop app | Yes cross-checked |
| Paid plan up to $20/month | Yes. $11–20/month |
| Paid plan up to $40/month | Yes. $11–20/month |
| Any phone client | Yes cross-checked |
Filter receipts
Desktop OS: macOS: Yes cross-checked
Desktop OS: macOS
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Desktop os: Stable 0.7.2 ships macOS Apple Silicon and Intel, Windows x64, and Linux x64 and arm64. The desktop app needs a desktop session; minimum OS versions were not established.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Notifications: Notification settings control alerts for finished threads and requests, including system and push.
- Cross device continuity: Continue on moves a thread between cloud and devices with branch, commits and uncommitted changes; running processes and ignored files stay behind. Not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
Desktop OS: Windows: Yes cross-checked
Desktop OS: Windows
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Desktop os: Stable 0.7.2 ships macOS Apple Silicon and Intel, Windows x64, and Linux x64 and arm64. The desktop app needs a desktop session; minimum OS versions were not established.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Notifications: Notification settings control alerts for finished threads and requests, including system and push.
- Cross device continuity: Continue on moves a thread between cloud and devices with branch, commits and uncommitted changes; running processes and ignored files stay behind. Not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
Desktop OS: Linux: Yes cross-checked
Desktop OS: Linux
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Desktop os: Stable 0.7.2 ships macOS Apple Silicon and Intel, Windows x64, and Linux x64 and arm64. The desktop app needs a desktop session; minimum OS versions were not established.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Notifications: Notification settings control alerts for finished threads and requests, including system and push.
- Cross device continuity: Continue on moves a thread between cloud and devices with branch, commits and uncommitted changes; running processes and ignored files stay behind. Not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
Web client: Yes cross-checked
Web client
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Browser: The web app starts agent threads.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Threads accept web, Slack, Linear and API prompts and carry hosted PR work.
Phone access: Native iOS: No cross-checked
Phone access: Native iOS
Typed value: No.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Mobile apps: Docs state there is no mobile app; phones use capy.ai in the browser.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Phone access: Native Android: No cross-checked
Phone access: Native Android
Typed value: No.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Mobile apps: Docs state there is no mobile app; phones use capy.ai in the browser.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Phone access: Phone browser: Yes cross-checked
Phone access: Phone browser
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Desktop os: Stable 0.7.2 ships macOS Apple Silicon and Intel, Windows x64, and Linux x64 and arm64. The desktop app needs a desktop session; minimum OS versions were not established.
- Mobile apps: Docs state there is no mobile app; phones use capy.ai in the browser.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Notifications: Notification settings control alerts for finished threads and requests, including system and push.
- Cross device continuity: Continue on moves a thread between cloud and devices with branch, commits and uncommitted changes; running processes and ignored files stay behind. Not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
CLI: Yes cross-checked
CLI
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Cli: capy dock is a headless device runner, not a task-control CLI.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: capy dock pairs a headless macOS, Linux or Windows machine as a device for threads.
Coding agents: Claude Code: Yes
Coding agents: Claude Code
Typed value: Yes.
Capy runs its own agent plus a Claude Code mode, with a wide provider catalog; other native harnesses were not found.
- Harnesses: Capy runs its own agent; Claude Code mode runs on the Claude Agent SDK with Capy tools and subagents. Other external harnesses were not found.
- Model providers: Anthropic, OpenAI, Google, xAI and others, plus gateways, coding plans and custom OpenAI-compatible endpoints.
- Session integration: Desktop can import selected local Codex and Claude Code sessions; Claude Code mode embeds the Agent SDK.
- Mixed harness operation: A thread picks Capy agent or Claude Code mode by model tab; each subagent can use its own model or connected subscription. Not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
Use my subscription: Yes cross-checked
Use my subscription
Typed value: Yes.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Where agents run: On your computer: Yes cross-checked
Where agents run: On your computer
Typed value: Yes.
Cloud VMs, desktop devices, Dock machines and movable threads are documented; own-cloud is Enterprise by agreement.
- Local: Desktop runs threads in a folder on the computer, with no machine-time charge.
- Hybrid: Continue on moves a thread between cloud and devices, and a Captain can run on either.
- Client disconnect: Cloud threads continue after the laptop closes; device threads wait up to 24 hours while offline.
- Other device continuation: Any browser can start, follow and reply while the device is online.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
Where agents run: On your own server: Yes cross-checked
Where agents run: On your own server
Typed value: Yes.
Cloud VMs, desktop devices, Dock machines and movable threads are documented; own-cloud is Enterprise by agreement.
- Byo remote: capy dock turns a user-run server or VM into a device.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: capy dock pairs a headless macOS, Linux or Windows machine as a device for threads.
Where agents run: In the vendor's cloud: Yes cross-checked
Where agents run: In the vendor's cloud
Typed value: Yes.
Cloud VMs, desktop devices, Dock machines and movable threads are documented; own-cloud is Enterprise by agreement.
- Managed cloud: Hosted thread agents execute commands in managed machines.
- Provisioning: Capy provisions a cloud VM from project environment setup.
- Payer: Cloud machine time draws credits even when model usage uses a subscription.
- Resources: 1 vCPU/4 GB through 16 vCPU/64 GB, gated by plan; Big Guy, 16 vCPU/128 GB, is on request.
- Region: US East (New York).
- Expiry: A machine sleeps about 2 minutes after work stops; one asleep for 30 days is deleted.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
Workspace isolation: Git worktree: Yes cross-checked
Workspace isolation: Git worktree
Typed value: Yes.
Cloud threads use Firecracker microVMs; device threads use worktrees and are explicitly not sandboxed.
- Checkout: Fresh writers start from a clean upstream checkout; a device thread can use Current checkout.
- Worktree: New worktree is the default on devices, under ~/.capy/worktrees.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
Workspace isolation: Container: No
Workspace isolation: Container
Typed value: No.
Cloud threads use Firecracker microVMs; device threads use worktrees and are explicitly not sandboxed.
- Container: Docker runs inside the VM. The product does not isolate work in containers.
- Vm: Firecracker microVM with its own kernel, per cloud thread.
- Filesystem: Shared children see parent uncommitted files; separate machines transfer files explicitly.
- Security boundary: Cloud machines hold no GitHub token and are isolated from other customers. Device threads have no sandbox.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
Workspace isolation: VM or microVM: Yes cross-checked
Workspace isolation: VM or microVM
Typed value: Yes.
Cloud threads use Firecracker microVMs; device threads use worktrees and are explicitly not sandboxed.
- Container: Docker runs inside the VM. The product does not isolate work in containers.
- Vm: Firecracker microVM with its own kernel, per cloud thread.
- Filesystem: Shared children see parent uncommitted files; separate machines transfer files explicitly.
- Security boundary: Cloud machines hold no GitHub token and are isolated from other customers. Device threads have no sandbox.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
Parallel runs: Yes cross-checked
Parallel runs
Typed value: Yes.
Concurrent child agents and thread-level activity aggregation are documented.
- Concurrent sessions: Subagents can work in parallel with separate conversations.
- Concurrent tasks: Parallel writers use disjoint files and separate PRs.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
Coordinator: Yes cross-checked
Coordinator
Typed value: Yes.
Subagents and Captain crews delegate with explicit prompts, per-child models and returned results.
- Roles: Parent and subagent, plus Captain and crew.
- Model harness effort: Each child can choose a model or connected subscription; Claude Code mode is a separate harness. Effort per child was not established.
- Messages: Children send summaries or blocked questions; the parent reads interim progress on demand.
- Results return: Final summaries wake the parent.
- Cross harness: Connected subscriptions and Claude Code mode allow mixed models in a tree. Not tested.
- Orchestration owner: The thread agent coordinates children; sibling threads can coordinate long-lived work.
- Activation: Ask the agent to split work; child drafts wait until started.
- Live progress: Interim progress is read on demand, not streamed into the parent automatically.
- Retained results: Crew reports stay in the Captain Drive; subagent results reach the parent.
- Followup: Parent may inspect a diff, patch it or send targeted revision feedback.
- Invocation: Subagents start from their explicit prompt without inheriting parent history.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
Native issue tracking: Yes cross-checked
Native issue tracking
Typed value: Yes.
Native Issues (GA in 0.7.0) add backlog, assignment, labels and search separate from threads.
- Create: Issues are created in the app, by threads or through the API.
- Edit: Issue fields update in the app and API; thread titles and subagent drafts are editable.
- Discuss: Issues have comments, and the owning thread wakes on a comment.
- Label: Issues carry labels; threads carry organization tags.
- Assign: Issues filter by assignee, including Assigned to me.
- Search: Issue search reads descriptions and comments; Cmd+K searches threads.
- Filter: Status, assignee, mentions, labels, due date and relations.
- Archive: Archive hides the thread and suppresses event-triggered waking.
- Reopen: A human message revives an archived web thread. Slack reopening uses an explicit unarchive command.
- Project epic views: Issue views cover For you, Assigned to me and Mentioned; sub-issues nest. An epic board was not established.
- Idea parking: An issue can exist without a running thread; threads pick up work from Issues.
- Explicit dispatch: Issues link to threads, and Captains take work from them. An explicit ready-to-dispatch control was not established.
- Task session separation: Issues are separate records linked to threads and pull requests.
- Metadata: Labels, due date, assignee, relations and states.
- Durable states: Issue states are configurable through the API, and threads have Working, Waiting, Failed and Idle.
- External tracker requirement: Web threads are usable without a Linear connection.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Mention starts a synced thread; files, channel mapping, stop and muted conversation controls; multiple people can reply.
-
Finding: Public API covers issues, labels, states, relations, comments and links to pull requests or threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
Unattended triggers: Yes cross-checked
Unattended triggers
Typed value: Yes.
Stored prompts support cron, GitHub/Slack events, webhooks and standing or fresh threads.
- Schedules: Five-field cron with IANA timezone and a five-minute minimum.
- Events: GitHub, Slack, incoming webhooks and on-demand triggers, plus agent timers and watches.
- Recurring tasks: An automation can create a new thread per run or reuse one standing thread.
- Scheduler location: Each run is a normal Capy thread, and the agent loop runs in the cloud.
- Retries: Webhook Idempotency-Key deduplicates retries of one delivery. General failed-run retry policy remains unverified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Stored prompts run on cron, GitHub/Slack events and webhooks; explicit principals, daily caps, idempotency and new or standing threads.
Shared prompting: Yes cross-checked
Shared prompting
Typed value: Yes.
Work is shared in an organization by default, and anyone with access can write in a thread; live ordering was not tested.
- Visibility: New work is visible to the organization unless made private.
- Coprompting: Anyone who can see a thread can write in it. Ordering was not tested.
- Permissions: Shared or private access per project, thread, Captain and folder.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
Team roles: Yes cross-checked
Team roles
Typed value: Yes.
Org roles, membership changes and private thread/project sharing are documented; enforcement was not tested.
- Organization roles: Owner/admin/member roles separate billing, administration and thread work.
- Membership: Invitations, role changes and member removal are documented.
- Plan scope: Roles appear on every plan; SSO and audit logging appear under Enterprise.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Owner/admin/member roles, membership changes, no seat charge and explicit service identities.
SSO: Yes cross-checked
SSO
Typed value: Yes.
Org roles, membership changes and private thread/project sharing are documented; enforcement was not tested.
- Sso: SSO/SAML is listed for Enterprise. Enforcement was not tested.
- Plan scope: Roles appear on every plan; SSO and audit logging appear under Enterprise.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Pull request workflow: Yes cross-checked
Pull request workflow
Typed value: Yes.
Branch/commit/push/PR creation, comments and CI follow-up are documented.
- Branches: Publishing from a base branch creates a fresh capy/ branch.
- Commits: The publishing tool commits and pushes.
- Pull requests: The agent opens PRs, including dependent stacked PRs.
- Comments: Human review/PR comments wake the thread; bot feedback is configurable.
- Ci response: Failure wakes repair and passing checks arrive as a consolidated result.
- Rebase: Update branch is offered; the owning thread can rebase stack layers on request.
- Conflicts: Conflict response defaults to React and wakes the owning thread.
- Product state: Pull request pages show stack, merge state and reviewers.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
Review interface: Yes cross-checked
Review interface
Typed value: Yes.
Threads retain PR/child results and Drive saves files beyond cloud-machine lifetime.
- Diffs: Threads show a growing diff; PR pages show completed diffs.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
Product license: Apache-2.0: No. Custom terms
Product license: Apache-2.0
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: MIT: No. Custom terms
Product license: MIT
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: GPL-3.0: No. Custom terms
Product license: GPL-3.0
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: GPL-3.0-or-later: No. Custom terms
Product license: GPL-3.0-or-later
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: AGPL-3.0: No. Custom terms
Product license: AGPL-3.0
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: AGPL-3.0-or-later: No. Custom terms
Product license: AGPL-3.0-or-later
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: Elastic License 2.0: No. Custom terms
Product license: Elastic License 2.0
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: Custom terms: Yes. Custom terms
Product license: Custom terms
Typed value: Yes. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Product license: Proprietary: No. Custom terms
Product license: Proprietary
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Free access: Yes cross-checked
Free access
Typed value: Yes.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Starting price: Free: Yes. Free cross-checked
Starting price: Free
Typed value: Yes. Free.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Starting price: Up to $20/month: No. Free cross-checked
Starting price: Up to $20/month
Typed value: No. Free.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Starting price: $21–40/month: No. Free cross-checked
Starting price: $21–40/month
Typed value: No. Free.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Starting price: $41–100/month: No. Free cross-checked
Starting price: $41–100/month
Typed value: No. Free.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Starting price: Over $100/month: No. Free cross-checked
Starting price: Over $100/month
Typed value: No. Free.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Starting price: Contact for pricing: No. Free cross-checked
Starting price: Contact for pricing
Typed value: No. Free.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Ongoing free: Threads on your own computer with a connected subscription or key need no plan or balance; provider cost remains.
- Paid only: Cloud machines need a plan or at least $5 of balance.
- Free scenarios: Local threads on BYO subscriptions and keys work on the free plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
Pricing model: Free, no paid plan: No
Pricing model: Free, no paid plan
Typed value: No.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Currency: Plan and compute rates are quoted in US dollars.
- Interval: Lite $20, Pro $100, Max $200 to $1,000 monthly; annual plans bill upfront at 20% less.
- Included usage: Lite $20, Pro $105, Max 110% of monthly price in credits.
- Credits: One org balance pays hosted models and cloud machine time.
- Model cost: Models have token rates on Capy credits; connected subscriptions/keys do not draw model charges from Capy balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
Pricing model: Free plan: Yes cross-checked
Pricing model: Free plan
Typed value: Yes.
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Paid evaluation: The paid Lite introduction renews at $20/month unless cancelled.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
Pricing model: Subscription: Yes cross-checked
Pricing model: Subscription
Typed value: Yes.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Pricing model: Per seat: No cross-checked
Pricing model: Per seat
Typed value: No.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Seats: Every plan has unlimited members and no per-seat price.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Owner/admin/member roles, membership changes, no seat charge and explicit service identities.
Pricing model: Usage-based: Yes cross-checked
Pricing model: Usage-based
Typed value: Yes.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Currency: Plan and compute rates are quoted in US dollars.
- Interval: Lite $20, Pro $100, Max $200 to $1,000 monthly; annual plans bill upfront at 20% less.
- Workspace fees: No charge for members or Team workspaces.
- Included usage: Lite $20, Pro $105, Max 110% of monthly price in credits.
- Credits: One org balance pays hosted models and cloud machine time.
- Overages: Add credits manually or through bounded auto-reload when usage exceeds the grant.
- Model cost: Models have token rates on Capy credits; connected subscriptions/keys do not draw model charges from Capy balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Plans, balance, auto-reload and usage monitoring by member, model and thread.
Pricing model: Per machine or compute: Yes cross-checked
Pricing model: Per machine or compute
Typed value: Yes.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Currency: Plan and compute rates are quoted in US dollars.
- Interval: Lite $20, Pro $100, Max $200 to $1,000 monthly; annual plans bill upfront at 20% less.
- Included usage: Lite $20, Pro $105, Max 110% of monthly price in credits.
- Credits: One org balance pays hosted models and cloud machine time.
- Model cost: Models have token rates on Capy credits; connected subscriptions/keys do not draw model charges from Capy balance.
- Compute cost: Cloud VM rates range from Small $0.10/hour through on-request Big Guy $3.20/hour; sleeping VMs cost nothing.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
Pricing model: Bring your own key: Yes cross-checked
Pricing model: Bring your own key
Typed value: Yes.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Currency: Plan and compute rates are quoted in US dollars.
- Interval: Lite $20, Pro $100, Max $200 to $1,000 monthly; annual plans bill upfront at 20% less.
- Included usage: Lite $20, Pro $105, Max 110% of monthly price in credits.
- Credits: One org balance pays hosted models and cloud machine time.
- Model cost: Models have token rates on Capy credits; connected subscriptions/keys do not draw model charges from Capy balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
Pricing model: One-time purchase: No
Pricing model: One-time purchase
Typed value: No.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Currency: Plan and compute rates are quoted in US dollars.
- Interval: Lite $20, Pro $100, Max $200 to $1,000 monthly; annual plans bill upfront at 20% less.
- Included usage: Lite $20, Pro $105, Max 110% of monthly price in credits.
- Credits: One org balance pays hosted models and cloud machine time.
- Model cost: Models have token rates on Capy credits; connected subscriptions/keys do not draw model charges from Capy balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
Paid plans from: No paid plan: No. $11–20/month
Paid plans from: No paid plan
Typed value: No. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Paid plans from: $10 or less/month: No. $11–20/month
Paid plans from: $10 or less/month
Typed value: No. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Paid plans from: $11–20/month: Yes. $11–20/month
Paid plans from: $11–20/month
Typed value: Yes. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Paid plans from: $21–40/month: No. $11–20/month
Paid plans from: $21–40/month
Typed value: No. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Paid plans from: $41–100/month: No. $11–20/month
Paid plans from: $41–100/month
Typed value: No. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Paid plans from: Over $100/month: No. $11–20/month
Paid plans from: Over $100/month
Typed value: No. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Paid plans from: Contact sales: No. $11–20/month
Paid plans from: Contact sales
Typed value: No. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Keeps going with laptop closed: Yes cross-checked
Keeps going with laptop closed
Typed value: Yes.
Cloud VMs, desktop devices, Dock machines and movable threads are documented; own-cloud is Enterprise by agreement.
- Local: Desktop runs threads in a folder on the computer, with no machine-time charge.
- Hybrid: Continue on moves a thread between cloud and devices, and a Captain can run on either.
- Client disconnect: Cloud threads continue after the laptop closes; device threads wait up to 24 hours while offline.
- Other device continuation: Any browser can start, follow and reply while the device is online.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
Open-source license: No. Custom terms
Open-source license
Typed value: No. Custom terms.
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Client license: The terms cover desktop pairing and the Capy software, with vendor-reserved intellectual property rights.
- Server license: The hosted platform is covered by the same vendor terms and retained rights.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
Any desktop app: Yes cross-checked
Any desktop app
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Desktop os: Stable 0.7.2 ships macOS Apple Silicon and Intel, Windows x64, and Linux x64 and arm64. The desktop app needs a desktop session; minimum OS versions were not established.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Notifications: Notification settings control alerts for finished threads and requests, including system and push.
- Cross device continuity: Continue on moves a thread between cloud and devices with branch, commits and uncommitted changes; running processes and ignored files stay behind. Not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
Paid plan up to $20/month: Yes. $11–20/month
Paid plan up to $20/month
Typed value: Yes. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Paid plan up to $40/month: Yes. $11–20/month
Paid plan up to $40/month
Typed value: Yes. $11–20/month.
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Trial: Monthly Lite introductory seven days cost $1 with a verified card, then $20 monthly.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
Any phone client: Yes cross-checked
Any phone client
Typed value: Yes.
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Desktop os: Stable 0.7.2 ships macOS Apple Silicon and Intel, Windows x64, and Linux x64 and arm64. The desktop app needs a desktop session; minimum OS versions were not established.
- Mobile apps: Docs state there is no mobile app; phones use capy.ai in the browser.
- Mobile browser: A phone browser can follow threads and review pull requests; on iPhone or iPad adding it to the Home Screen enables push. Review parity was not tested.
- Notifications: Notification settings control alerts for finished threads and requests, including system and push.
- Cross device continuity: Continue on moves a thread between cloud and devices with branch, commits and uncommitted changes; running processes and ignored files stay behind. Not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
All public findings
Findings may summarize a source. They are not presented as verbatim statements.
capy:ACC-01: Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app. cross-checked
capy:ACC-01
Web and desktop clients are documented, a headless Dock CLI hosts threads, and docs state there is no mobile app.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Threads accept web, Slack, Linear and API prompts and carry hosted PR work.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: capy dock pairs a headless macOS, Linux or Windows machine as a device for threads.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:ACC-02: Capy runs its own agent plus a Claude Code mode, with a wide provider catalog; other native harnesses were not found. cross-checked
capy:ACC-02
Capy runs its own agent plus a Claude Code mode, with a wide provider catalog; other native harnesses were not found.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
capy:ACC-04: Cloud GitHub onboarding, repoless projects, desktop folders and Dock pairing are documented. cross-checked
capy:ACC-04
Cloud GitHub onboarding, repoless projects, desktop folders and Dock pairing are documented.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Multi-repository projects with base branches, stable thread codes, GitHub-scoped clones and repoless projects.
-
Finding: Versioned per-repository setup scripts and services, repository setup file, snapshot restore and fallback boot, explicit command launch.
-
Finding: The desktop app starts threads with no project.
-
Finding: capy dock pairs a headless macOS, Linux or Windows machine as a device for threads.
-
Finding: GitHub.com only; GitHub Enterprise Server and other hosts unsupported.
capy:ACC-05: A public bearer-key API covers threads, Captains, automations, reviews, usage and issues; capy dock is a device runner. cross-checked
capy:ACC-05
A public bearer-key API covers threads, Captains, automations, reviews, usage and issues; capy dock is a device runner.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Stored prompts run on cron, GitHub/Slack events and webhooks; explicit principals, daily caps, idempotency and new or standing threads.
-
Finding: Bearer keys act as explicit principals; threads, messages, streams, automations, read-only subagents, reviews and usage.
-
Finding: capy dock pairs a headless macOS, Linux or Windows machine as a device for threads.
-
Finding: Public API covers issues, labels, states, relations, comments and links to pull requests or threads.
capy:ACC-06: MCP servers, marketplace plugins and portable repository instructions are documented.
capy:ACC-06
MCP servers, marketplace plugins and portable repository instructions are documented.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: AGENTS.md and CLAUDE.md fallback, nested and glob rules, organization/personal instructions; next-thread update semantics.
-
Finding: Org/project HTTP and local stdio MCP tools with shared/personal credentials; dedicated review agent never receives MCP tools.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
-
Finding: Marketplace and GitHub plugins bundle MCP servers and skills.
-
Finding: Threads and subagents call custom MCP tools over Streamable HTTP or local STDIO, with scoped connections.
capy:PLAN-01: Prompts, Slack, Linear delegation and native Issues feed durable threads. cross-checked
capy:PLAN-01
Prompts, Slack, Linear delegation and native Issues feed durable threads.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Stored prompts run on cron, GitHub/Slack events and webhooks; explicit principals, daily caps, idempotency and new or standing threads.
-
Finding: Mention starts a synced thread; files, channel mapping, stop and muted conversation controls; multiple people can reply.
-
Finding: Native Linear agent sessions map to Capy threads, stream plans/progress, preserve identifiers and explicit user links.
-
Finding: Public API covers issues, labels, states, relations, comments and links to pull requests or threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:PLAN-02: There is no plan mode; plans live in prompts, Captain notes and Drive, with no acceptance-criteria field. cross-checked
capy:PLAN-02
There is no plan mode; plans live in prompts, Captain notes and Drive, with no acceptance-criteria field.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Shared and personal persisted files outside Git, browser editing, version history and restore; older versions kept for 24 hours.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:PLAN-03: Issues add sub-issues and blockers, and threads can grow into Captains; blocker enforcement was not established. cross-checked
capy:PLAN-03
Issues add sub-issues and blockers, and threads can grow into Captains; blocker enforcement was not established.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Public API covers issues, labels, states, relations, comments and links to pull requests or threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:PLAN-04: Repository and Drive instructions, skills and saved files provide scoped shared context. cross-checked
capy:PLAN-04
Repository and Drive instructions, skills and saved files provide scoped shared context.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: AGENTS.md and CLAUDE.md fallback, nested and glob rules, organization/personal instructions; next-thread update semantics.
-
Finding: Built-in, repository and Drive skills load on demand; portable SKILL.md workflow files.
-
Finding: Shared and personal persisted files outside Git, browser editing, version history and restore; older versions kept for 24 hours.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:PLAN-05: Images, files and scoped Slack context are documented; architecture indexing and review enforcement are unverified. cross-checked
capy:PLAN-05
Images, files and scoped Slack context are documented; architecture indexing and review enforcement are unverified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Mention starts a synced thread; files, channel mapping, stop and muted conversation controls; multiple people can reply.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
capy:PLAN-06: Native Issues (GA in 0.7.0) add backlog, assignment, labels and search separate from threads. cross-checked
capy:PLAN-06
Native Issues (GA in 0.7.0) add backlog, assignment, labels and search separate from threads.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Mention starts a synced thread; files, channel mapping, stop and muted conversation controls; multiple people can reply.
-
Finding: Public API covers issues, labels, states, relations, comments and links to pull requests or threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:PLAN-07: Threads read, file, update and close shared issues, which are also in the public API. cross-checked
capy:PLAN-07
Threads read, file, update and close shared issues, which are also in the public API.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Bearer keys act as explicit principals; threads, messages, streams, automations, read-only subagents, reviews and usage.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
-
Finding: Public API covers issues, labels, states, relations, comments and links to pull requests or threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:PLAN-08: Repository specification files and Drive support an instruction-based spec workflow with human and agent edits and short version history. Dedicated specification links and drift detection remain unknown. cross-checked
capy:PLAN-08
Repository specification files and Drive support an instruction-based spec workflow with human and agent edits and short version history. Dedicated specification links and drift detection remain unknown.
- This is a file and instruction workflow; dedicated specification governance and drift checks were not established.
-
Finding: Drive stores files across threads. People and agents edit them, with automatic saving and 24-hour version history naming originating threads.
-
Finding: Users can request a repository spec before changes. Stop run cancels waiting timers and background commands but leaves subagents running; shared-machine side chats compete for resources.
capy:COORD-01: Concurrent child agents and thread-level activity aggregation are documented. cross-checked
capy:COORD-01
Concurrent child agents and thread-level activity aggregation are documented.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
capy:COORD-02: Issues hold assignees, and Captains and parents dispatch work; exclusive task claims were not established. cross-checked
capy:COORD-02
Issues hold assignees, and Captains and parents dispatch work; exclusive task claims were not established.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Mention starts a synced thread; files, channel mapping, stop and muted conversation controls; multiple people can reply.
-
Finding: Native Linear agent sessions map to Capy threads, stream plans/progress, preserve identifiers and explicit user links.
-
Finding: Bearer keys act as explicit principals; threads, messages, streams, automations, read-only subagents, reviews and usage.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:COORD-03: Subagents and Captain crews delegate with explicit prompts, per-child models and returned results. cross-checked
capy:COORD-03
Subagents and Captain crews delegate with explicit prompts, per-child models and returned results.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
capy:COORD-04: Stored prompts support cron, GitHub/Slack events, webhooks and standing or fresh threads. cross-checked
capy:COORD-04
Stored prompts support cron, GitHub/Slack events, webhooks and standing or fresh threads.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Stored prompts run on cron, GitHub/Slack events and webhooks; explicit principals, daily caps, idempotency and new or standing threads.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
capy:COORD-05: Steer, interrupt and queued messages coexist with Stop run, Needs you and Failed status. cross-checked
capy:COORD-05
Steer, interrupt and queued messages coexist with Stop run, Needs you and Failed status.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:COORD-06: Status, Needs you, Captain digests and issue blockers support an overview; no fleet fixture was run. cross-checked
capy:COORD-06
Status, Needs you, Captain digests and issue blockers support an overview; no fleet fixture was run.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:COORD-07: A thread can stay simple or be promoted to a Captain that directs other threads. cross-checked
capy:COORD-07
A thread can stay simple or be promoted to a Captain that directs other threads.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:RUN-01: Cloud VMs, desktop devices, Dock machines and movable threads are documented; own-cloud is Enterprise by agreement. cross-checked
capy:RUN-01
Cloud VMs, desktop devices, Dock machines and movable threads are documented; own-cloud is Enterprise by agreement.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Versioned per-repository setup scripts and services, repository setup file, snapshot restore and fallback boot, explicit command launch.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: capy dock pairs a headless macOS, Linux or Windows machine as a device for threads.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:RUN-02: Cloud threads use Firecracker microVMs; device threads use worktrees and are explicitly not sandboxed. cross-checked
capy:RUN-02
Cloud threads use Firecracker microVMs; device threads use worktrees and are explicitly not sandboxed.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: Encrypted variables injected at boot; agent sees names only, but code on the machine can read values.
capy:RUN-03: Cloud machine identity survives backing replacement; filesystem and process continuity differ. cross-checked
capy:RUN-03
Cloud machine identity survives backing replacement; filesystem and process continuity differ.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: Replacement keeps pushed work; failed setup still starts the machine with a warning.
capy:RUN-04: Versioned setup, snapshots and startup entries prepare and rebuild managed environments. cross-checked
capy:RUN-04
Versioned setup, snapshots and startup entries prepare and rebuild managed environments.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Versioned per-repository setup scripts and services, repository setup file, snapshot restore and fallback boot, explicit command launch.
-
Finding: Built-in, repository and Drive skills load on demand; portable SKILL.md workflow files.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Replacement keeps pushed work; failed setup still starts the machine with a warning.
-
Finding: Named test, lint and typecheck commands run on demand. Snapshots cache installed setup; restored checkouts refresh only after HEAD moves.
capy:RUN-05: Credits and connection exhaustion have visible stop/recovery behavior; broad failure guarantees are unverified. cross-checked
capy:RUN-05
Credits and connection exhaustion have visible stop/recovery behavior; broad failure guarantees are unverified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Versioned per-repository setup scripts and services, repository setup file, snapshot restore and fallback boot, explicit command launch.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Plans, balance, auto-reload and usage monitoring by member, model and thread.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
capy:RUN-06: Live progress and spend/token breakdowns are documented; aggregate metrics were not tested. cross-checked
capy:RUN-06
Live progress and spend/token breakdowns are documented; aggregate metrics were not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Stored prompts run on cron, GitHub/Slack events and webhooks; explicit principals, daily caps, idempotency and new or standing threads.
-
Finding: Plans, balance, auto-reload and usage monitoring by member, model and thread.
-
Finding: Bearer keys act as explicit principals; threads, messages, streams, automations, read-only subagents, reviews and usage.
-
Finding: Windowed organization model/image/VM spend and token totals with member/model breakdowns; no per-child field established.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:RUN-07: Per-thread cost, usage splits and exports exist; child, epic and provider reconciliation were not established. cross-checked
capy:RUN-07
Per-thread cost, usage splits and exports exist; child, epic and provider reconciliation were not established.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, balance, auto-reload and usage monitoring by member, model and thread.
-
Finding: Bearer keys act as explicit principals; threads, messages, streams, automations, read-only subagents, reviews and usage.
-
Finding: Windowed organization model/image/VM spend and token totals with member/model breakdowns; no per-child field established.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:RUN-08: Plan-specific cloud machine and concurrent-RAM limits are documented; compute-aware arbitration and verification reuse remain unverified. cross-checked
capy:RUN-08
Plan-specific cloud machine and concurrent-RAM limits are documented; compute-aware arbitration and verification reuse remain unverified.
- Only published capacity configuration was checked; no runtime scheduling or test-reuse scenario was run.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Users can request a repository spec before changes. Stop run cancels waiting timers and background commands but leaves subagents running; shared-machine side chats compete for resources.
capy:TEAM-01: Work is shared in an organization by default, and anyone with access can write in a thread; live ordering was not tested. cross-checked
capy:TEAM-01
Work is shared in an organization by default, and anyone with access can write in a thread; live ordering was not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Owner/admin/member roles, membership changes, no seat charge and explicit service identities.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Mention starts a synced thread; files, channel mapping, stop and muted conversation controls; multiple people can reply.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:TEAM-02: Org roles, membership changes and private thread/project sharing are documented; enforcement was not tested. cross-checked
capy:TEAM-02
Org roles, membership changes and private thread/project sharing are documented; enforcement was not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Owner/admin/member roles, membership changes, no seat charge and explicit service identities.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
-
Finding: Email invitations and verified-domain joining provision members. All plans include unlimited members without seat charges.
capy:TEAM-03: Device access has approve, deny and always-allow controls; the merge rule is an instruction, and cloud commands have no per-command approval found.
capy:TEAM-03
Device access has approve, deny and always-allow controls; the merge rule is an instruction, and cloud commands have no per-command approval found.
- Some atomic questions remain unknown.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
-
Finding: capy dock pairs a headless macOS, Linux or Windows machine as a device for threads.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
-
Finding: Encrypted variables injected at boot; agent sees names only, but code on the machine can read values.
-
Finding: Cloud machines join a tailnet as ephemeral tagged nodes. User ACLs control private service access; machine auth keys expire after five minutes and are absent from snapshots.
capy:TEAM-04: Threads retain conversation/activity and Drive retains attributed file versions; immutable audit guarantees are unverified. cross-checked
capy:TEAM-04
Threads retain conversation/activity and Drive retains attributed file versions; immutable audit guarantees are unverified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Shared and personal persisted files outside Git, browser editing, version history and restore; older versions kept for 24 hours.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:TEAM-05: PR pages expose diffs, checks and findings, with review comments sent back to the agent. cross-checked
capy:TEAM-05
PR pages expose diffs, checks and findings, with review comments sent back to the agent.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:TEAM-06: Ownership transfer exists and shared threads are open to teammates; runtime takeover was not tested. cross-checked
capy:TEAM-06
Ownership transfer exists and shared threads are open to teammates; runtime takeover was not tested.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Mention starts a synced thread; files, channel mapping, stop and muted conversation controls; multiple people can reply.
-
Finding: 0.4.6 release, dated 9 October 2026, added private sharing and ownership transfer.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Wake events, steer/queue/interrupt, side chats, forks, subagents, crew threads.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:SHIP-01: Threads retain PR/child results and Drive saves files beyond cloud-machine lifetime. cross-checked
capy:SHIP-01
Threads retain PR/child results and Drive saves files beyond cloud-machine lifetime.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Agent-created editable drafts, three child levels, shared read-only or fresh writer machines, sequential stacks, result messages and per-child models.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Shared and personal persisted files outside Git, browser editing, version history and restore; older versions kept for 24 hours.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
-
Finding: Pull request pages show Vercel deployment status and preview links.
capy:SHIP-02: CI failures trigger repair and dedicated reviews triage findings; spec gates and enforced branch policy are unverified. cross-checked
capy:SHIP-02
CI failures trigger repair and dedicated reviews triage findings; spec gates and enforced branch policy are unverified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Named test, lint and typecheck commands run on demand. Snapshots cache installed setup; restored checkouts refresh only after HEAD moves.
capy:SHIP-03: Branch/commit/push/PR creation, comments and CI follow-up are documented. cross-checked
capy:SHIP-03
Branch/commit/push/PR creation, comments and CI follow-up are documented.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
capy:SHIP-04: The documented merge rule requires a specific explicit ask; technical authority enforcement is unverified. cross-checked
capy:SHIP-04
The documented merge rule requires a specific explicit ask; technical authority enforcement is unverified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:SHIP-05: PR links and merge events connect results to durable threads; rollback mechanics are unverified. cross-checked
capy:SHIP-05
PR links and merge events connect results to durable threads; rollback mechanics are unverified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: GitHub App and project setup, live todos and diffs, PR creation and event-driven CI/review follow-up.
-
Finding: Status Working, Waiting, Failed or Idle, with a separate Needs you flag; steer, interrupt and queue.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Pull request pages show Vercel deployment status and preview links.
capy:SHIP-06: Pull request pages, review guides, findings and stacks support batch review; a six-task queue was not run.
capy:SHIP-06
Pull request pages, review guides, findings and stacks support batch review; a six-task queue was not run.
- Some atomic questions remain unknown.
-
Finding: Branch creation at publish, commit/push/PR tool, CI failures and human comments wake threads; stacked PR bases and last-push ownership.
-
Finding: Dedicated reviewer checks exact PR head, posts inline GitHub findings, reruns on pushes, carries open findings and respects triage.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
-
Finding: Long-lived agent directing a crew of threads, with digest, Drive plan and owner-scoped permissions.
capy:OWN-01: The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
capy:OWN-01
The Capy application and hosted service use vendor-authored terms. Capy and its licensors retain rights in the software, interfaces and platform, and restrict modification, distribution and reverse engineering without permission.
- Only the named documented or inspected parts are established; other atomic questions retain their recorded gaps.
-
Finding: The vendor terms cover the Capy platform, desktop device pairing and agents. Capy and licensors retain rights in the service software and restrict modification, redistribution and reverse engineering without permission.
capy:OWN-02: Threads stay in Capy cloud even on devices; own-cloud is Enterprise by agreement, and export is through the API. cross-checked
capy:OWN-02
Threads stay in Capy cloud even on devices; own-cloud is Enterprise by agreement, and export is through the API.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, balance, auto-reload and usage monitoring by member, model and thread.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
-
Finding: Local threads in worktrees or current checkout; offline wait up to 24 hours; Continue on moves work between machines.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:OWN-03: Capy documents storage, retention, deletion, subprocessors and telemetry, and provider terms apply on your own keys. cross-checked
capy:OWN-03
Capy documents storage, retention, deletion, subprocessors and telemetry, and provider terms apply on your own keys.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Vendor states SOC 2 Type II with report in Trust Center; provider terms govern provider processing; enterprise SSO and audit/export options by agreement.
-
Finding: Page lists subprocessors including Anthropic, OpenAI, Google, AWS, Cloudflare, Vercel, Stripe, Sentry and PostHog.
capy:OWN-04: Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs. cross-checked
capy:OWN-04
Monthly plans grant shared credits, with no member fee; cloud compute and model usage have separate costs.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Owner/admin/member roles, membership changes, no seat charge and explicit service identities.
-
Finding: Plans, balance, auto-reload and usage monitoring by member, model and thread.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
capy:OWN-05: Only the fixed and variable rate parts of a scenario can be priced from public sources.
capy:OWN-05
Only the fixed and variable rate parts of a scenario can be priced from public sources.
- Some atomic questions remain unknown.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:OWN-06: Public downloads and dated releases are available; support response commitments were not verified. cross-checked
capy:OWN-06
Public downloads and dated releases are available; support response commitments were not verified.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Bearer keys act as explicit principals; threads, messages, streams, automations, read-only subagents, reviews and usage.
-
Finding: Stable 0.7.2 for macOS, Windows and Linux; local session import; FAQ says a Claude subscription cannot be connected.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
-
Finding: Dated release feed from 0.4.2 through 0.7.2.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
-
Finding: iOS and Android cloud phones attach to a thread on Pro, Max or Enterprise.
-
Finding: In-app chat, email support@capy.ai, Discord and X.
-
Finding: Captains and Issues GA on 31 October 2026; issues hold blockers, relations, labels, due dates and sub-issues; threads file, update and close issues; Issues are in the public API.
capy:OWN-07: Capy publicly states SOC 2 Type II and provides a Trust Center; the report scope and period were not inspected. cross-checked
capy:OWN-07
Capy publicly states SOC 2 Type II and provides a Trust Center; the report scope and period were not inspected.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Cloud credential proxy and scoped encrypted secrets; explicit merge request rule; encrypted backend history and machine retention.
-
Finding: Vendor states SOC 2 Type II with report in Trust Center; provider terms govern provider processing; enterprise SSO and audit/export options by agreement.
-
Finding: Trust Center returned HTTP 200; report contents were not accessed.
capy:OWN-08: Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance. cross-checked
capy:OWN-08
Local threads on your own subscriptions or keys need no plan; cloud threads need a plan or balance.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Managed Ubuntu VMs; filesystem survives sleep but processes do not; deletion after 30 idle days; cross-machine file transfer.
-
Finding: Shared and personal persisted files outside Git, browser editing, version history and restore; older versions kept for 24 hours.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.
capy:OWN-09: Public plan, model and hourly machine rates exist, with cancel/top-up controls and custom enterprise terms. cross-checked
capy:OWN-09
Public plan, model and hourly machine rates exist, with cancel/top-up controls and custom enterprise terms.
- Remaining atomic behavior is unverified; this is a documentation pass, not a hands-on test.
-
Finding: Plans, model rates, subscriptions, connections and credit rules; connections work on every plan.
-
Finding: Plans, balance, auto-reload and usage monitoring by member, model and thread.
-
Finding: Current plans list unlimited members, Lite $20 monthly, Pro $100, Max $200+ and Enterprise custom; paid $1 seven-day Lite introduction.
-
Finding: Vendor states SOC 2 Type II with report in Trust Center; provider terms govern provider processing; enterprise SSO and audit/export options by agreement.
-
Finding: Connections work on every plan; no mobile app; US East machines; no export button; no plan mode.