bb
Open-source desktop app for Mac, Windows and Linux. Agents run on your computer or on your own server.
Product websitegithub.com Compare with others Missing or wrong data?
- Desktop app
- Mac, Windows and Linux
- Browser
- Yes
- Phone
- IPhone, Android and phone browser
- Command line
- Yes
- Agents
- Claude Code, Codex, Cursor Agent, OpenCode, Pi and Grok
- Your subscription
- Works with your own plan
- Agents run
- On your computer and on your own server
- Laptop closed
- Keeps going
- Isolation
- Worktree and Container
- Parallel agents
- Yes
- Scheduled runs
- Yes
- Coordinator agent
- Yes
- Built-in issues
- Yes
- Pull requests
- Yes
- Review screen
- Yes
- Team
- Not listed
- Self-hosting
- Yes
- Licence
- MIT
- Price from
- Free
- How it charges
- No paid plan and bring your own key
- Paid plans from
- Not listed
Measured by ADEMark
ADEMark ran the same tests on 15 ADEs. Results for bb 0.49.0, 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
- 2.25 s5th of 12median of 6 runs, 2.23 s to 2.29 s · Bare CLI 0.69 s
- New thread, ready to type
- 244 ms5th of 10median of 9 runs, 227 ms to 261 ms · Bare CLI 375 ms
- First word on screen
- 903 ms9th of 11median of 9 runs, 871 ms to 971 ms · Bare CLI 220 ms
- Typing lag, 5 agents streaming
- 62 ms5th of 111 run · Bare CLI 79 ms
- Open a 2,000-message thread
- 352 ms11th of 11median of 5 runs, 335 ms to 366 ms · Bare CLI 203 ms
Weight
- Memory at rest, 5 threads
- 1.49 GB9th of 121 run · Bare CLI 0.99 GB
- Idle CPU
- 2.18%8th of 131 run · Bare CLI 0.37%
- Memory under load
- 2.62 GB8th of 101 run · Bare CLI 1.86 GB
Tokens
- Tokens to answer “hi”
- 24.5K tokens2nd of 13median of 3 runs, 24.5K tokens to 24.5K tokens · Bare CLI 38.2K tokens
- Tokens for one real question
- 81.8K tokens2nd of 131 run · Bare CLI 168.3K tokens
Build
- Team build: hidden tests passed
- 0 of 53rd of 41 run
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 2026-11-08
- Findings
- 38 published · 34 cross-checked by a second agent · read them all
- Sources
- 27 public pages on 3 sites: github.com, getbb.app, cdn.agentclientprotocol.com
- Screenshot
- github.com, retrieved · first party repository · shows appearance, not capability
- Data
- /products/bb.json · .md
- Method
- Same rules for all 103 researched ADEs. How we check · Report wrong data
What it does
Where you use it
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
Source and scope for Where you use itAgents it runs
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
Source and scope for Agents it runsAccounts and subscriptions
Codex accepts ChatGPT login, while Claude uses the signed-in CLI and reports its plan. Codex also accepts an OpenAI API key.
Source and scope for Accounts and subscriptionsWhere agents run
Default launcher starts server and local host daemon on the current machine. Enrolled execution machines run provider processes remotely under one bb server. A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work. Optional Modal machines use the operator own provider account; there is no bb-vendor-operated execution service in the scoped source.
Source and scope for Where agents runHow work is kept apart
Managed worktrees separate checkout directories/branches. The optional official Modal plugin adds image-based container sandboxes in the operator account; no selectable VM/microVM environment backend ships in this revision.
Source and scope for How work is kept apartParallel work
Managed worktrees allow concurrent threads without changing the main checkout.
Source and scope for Parallel workScheduled runs
Automations support repeating cron/timezone and one-shot schedules. Automation scripts execute on the bb server host; agent runs use installed providers.
Source and scope for Scheduled runsGit and pull requests
Managed worktree environments create separate branches for parallel work. GitHub PR status plugin exposes PR state, Mark ready and merge actions.
Source and scope for Git and pull requestsLicence and source
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
Source and scope for Licence and sourcePrice
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
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 | Yes cross-checked |
| Phone access: Native Android | Yes cross-checked |
| Phone access: Phone browser | Yes cross-checked |
| CLI | Yes cross-checked |
| Coding agents: Claude Code | Yes cross-checked |
| Coding agents: Codex | Yes cross-checked |
| Coding agents: Cursor Agent | Yes cross-checked |
| Coding agents: Gemini CLI | Unknown |
| Coding agents: OpenCode | Yes cross-checked |
| Coding agents: Pi | Yes cross-checked |
| Coding agents: Grok | Yes cross-checked |
| Coding agents: Aider | Unknown |
| Coding agents: Antigravity | Unknown |
| Use my subscription | Yes cross-checked |
| Where agents run: On your computer | Yes |
| Where agents run: On your own server | Yes |
| Where agents run: In the vendor's cloud | No |
| Workspace isolation: Git worktree | Yes |
| Workspace isolation: Container | Yes |
| Workspace isolation: VM or microVM | No |
| Parallel runs | Yes cross-checked |
| Coordinator | Yes cross-checked |
| Native issue tracking | Yes cross-checked |
| Unattended triggers | Yes cross-checked |
| Shared prompting | Unknown |
| Team roles | No |
| SSO | No |
| Pull request workflow | Yes cross-checked |
| Review interface | Yes cross-checked |
| Product license: Apache-2.0 | No. MIT cross-checked |
| Product license: MIT | Yes. MIT cross-checked |
| Product license: GPL-3.0 | No. MIT cross-checked |
| Product license: GPL-3.0-or-later | No. MIT cross-checked |
| Product license: AGPL-3.0 | No. MIT cross-checked |
| Product license: AGPL-3.0-or-later | No. MIT cross-checked |
| Product license: Elastic License 2.0 | No. MIT cross-checked |
| Product license: Custom terms | No. MIT cross-checked |
| Product license: Proprietary | No. MIT cross-checked |
| Self-hosting | Yes cross-checked |
| Telemetry opt-out | Yes cross-checked |
| 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 | Yes |
| Pricing model: Free plan | No |
| Pricing model: Subscription | No |
| Pricing model: Per seat | No |
| Pricing model: Usage-based | No |
| Pricing model: Per machine or compute | No |
| Pricing model: Bring your own key | Yes |
| Pricing model: One-time purchase | No |
| Paid plans from: No paid plan | Unknown |
| Paid plans from: $10 or less/month | Unknown |
| Paid plans from: $11–20/month | Unknown |
| Paid plans from: $21–40/month | Unknown |
| Paid plans from: $41–100/month | Unknown |
| Paid plans from: Over $100/month | Unknown |
| Paid plans from: Contact sales | Unknown |
| Lifecycle status: Active | Unknown |
| Lifecycle status: Sunsetting | Unknown |
| Keeps going with laptop closed | Yes |
| Open-source license | Yes. MIT cross-checked |
| Any desktop app | Yes cross-checked |
| Paid plan up to $20/month | Unknown |
| Paid plan up to $40/month | Unknown |
| Any phone client | Yes cross-checked |
Filter receipts
Desktop OS: macOS: Yes cross-checked
Desktop OS: macOS
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Desktop os: Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
Desktop OS: Windows: Yes cross-checked
Desktop OS: Windows
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Desktop os: Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
Desktop OS: Linux: Yes cross-checked
Desktop OS: Linux
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Desktop os: Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
Web client: Yes cross-checked
Web client
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Browser: The server serves a local or remote browser app for live thread inspection and steering.
- Mobile browser: Phones and tablets can use a remote browser without the optional local editor helper.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
Phone access: Native iOS: Yes cross-checked
Phone access: Native iOS
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Mobile apps: iOS TestFlight and Android APK clients connect to a bb server.
- Notifications: The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile, and a plain-HTTP direct profile cannot register.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: iOS TestFlight and Android APK clients connect to the server. Phone apps are control clients with documented surface limitations, not execution hosts.
Phone access: Native Android: Yes cross-checked
Phone access: Native Android
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Mobile apps: iOS TestFlight and Android APK clients connect to a bb server.
- Notifications: The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile, and a plain-HTTP direct profile cannot register.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: iOS TestFlight and Android APK clients connect to the server. Phone apps are control clients with documented surface limitations, not execution hosts.
Phone access: Phone browser: Yes cross-checked
Phone access: Phone browser
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Browser: The server serves a local or remote browser app for live thread inspection and steering.
- Mobile browser: Phones and tablets can use a remote browser without the optional local editor helper.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
CLI: Yes cross-checked
CLI
Typed value: Yes.
bb CLI exposes thread and project operations; plugin commands extend it. The central server exposes HTTP API and WebSocket updates. Direct API is unauthenticated behind the documented trusted-network boundary. The npm package exports a Node SDK including thread spawn. The MCP plugin exposes CLI search/read/run for remote or local clients.
- Cli: bb CLI exposes thread and project operations; plugin commands extend it.
- Sdk: The npm package exports a Node SDK including thread spawn.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Coding agents: Claude Code: Yes cross-checked
Coding agents: Claude Code
Typed value: Yes.
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Harnesses: Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented.
- Mixed harness operation: Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Coding agents: Codex: Yes cross-checked
Coding agents: Codex
Typed value: Yes.
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Harnesses: Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented.
- Mixed harness operation: Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Coding agents: Cursor Agent: Yes cross-checked
Coding agents: Cursor Agent
Typed value: Yes.
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Harnesses: Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented.
- Mixed harness operation: Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Coding agents: OpenCode: Yes cross-checked
Coding agents: OpenCode
Typed value: Yes.
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Harnesses: Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented.
- Mixed harness operation: Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Coding agents: Pi: Yes cross-checked
Coding agents: Pi
Typed value: Yes.
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Harnesses: Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented.
- Mixed harness operation: Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Coding agents: Grok: Yes cross-checked
Coding agents: Grok
Typed value: Yes.
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Harnesses: Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented.
- Mixed harness operation: Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Use my subscription: Yes cross-checked
Use my subscription
Typed value: Yes.
Codex accepts ChatGPT login, while Claude uses the signed-in CLI and reports its plan. Codex also accepts an OpenAI API key.
- Agent subscriptions: Codex accepts ChatGPT login, while Claude uses the signed-in CLI and reports its plan.
- Api accounts: Codex also accepts an OpenAI API key.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/provider-claude-code/PLUGIN_OVERVIEW.md
Finding: Claude provider runs the installed signed-in Claude CLI and reads plan/usage; permission modes and optional disabling of provider-native subagents are documented.
-
bb plugins/provider-codex/PLUGIN_OVERVIEW.md
Finding: Codex supports ChatGPT login or API key, shows ChatGPT usage limits and can disable provider-native subagents so delegation uses bb.
Where agents run: On your computer: Yes
Where agents run: On your computer
Typed value: Yes.
Default launcher starts server and local host daemon on the current machine. Enrolled execution machines run provider processes remotely under one bb server. A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work. Optional Modal machines use the operator own provider account; there is no bb-vendor-operated execution service in the scoped source.
- Local: Default launcher starts server and local host daemon on the current machine.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
Where agents run: On your own server: Yes
Where agents run: On your own server
Typed value: Yes.
Default launcher starts server and local host daemon on the current machine. Enrolled execution machines run provider processes remotely under one bb server. A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work. Optional Modal machines use the operator own provider account; there is no bb-vendor-operated execution service in the scoped source.
- Byo remote: Enrolled execution machines run provider processes remotely under one bb server.
- Client disconnect: A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
Where agents run: In the vendor's cloud: No
Where agents run: In the vendor's cloud
Typed value: No.
Default launcher starts server and local host daemon on the current machine. Enrolled execution machines run provider processes remotely under one bb server. A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work. Optional Modal machines use the operator own provider account; there is no bb-vendor-operated execution service in the scoped source.
- Managed cloud: The optional hosted sandbox backend uses operator Modal credentials; other execution machines are user-enrolled hosts.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/environment-modal-sandbox/README.md
Finding: The optional official Modal sandbox plugin builds a named Dockerfile image, creates a reusable sandbox machine in the user Modal account, enrolls its daemon and clones projects. Operator Modal credentials pay provider infrastructure; bb supplies no vendor-operated execution service. The shipped environment backend set has host checkout/worktree or Modal image-based sandbox, with no selectable VM/microVM backend.
Workspace isolation: Git worktree: Yes
Workspace isolation: Git worktree
Typed value: Yes.
Managed worktrees separate checkout directories/branches. The optional official Modal plugin adds image-based container sandboxes in the operator account; no selectable VM/microVM environment backend ships in this revision.
- Worktree: Managed Git worktrees create separate working directories and branches per environment.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Managed worktrees create separate branches, copy selected untracked files and run setup/teardown hooks. Setup failures prevent thread start; this is file separation, not a claimed security sandbox.
Workspace isolation: Container: Yes
Workspace isolation: Container
Typed value: Yes.
Managed worktrees separate checkout directories/branches. The optional official Modal plugin adds image-based container sandboxes in the operator account; no selectable VM/microVM environment backend ships in this revision.
- Container: Official Modal sandbox plugin builds Dockerfile images and creates reusable sandbox machines from them.
- Vm: Complete shipped environment provider set has host checkouts/worktrees and Modal image-based sandboxes, without a VM/microVM backend.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/environment-modal-sandbox/README.md
Finding: The optional official Modal sandbox plugin builds a named Dockerfile image, creates a reusable sandbox machine in the user Modal account, enrolls its daemon and clones projects. Operator Modal credentials pay provider infrastructure; bb supplies no vendor-operated execution service. The shipped environment backend set has host checkout/worktree or Modal image-based sandbox, with no selectable VM/microVM backend.
Workspace isolation: VM or microVM: No
Workspace isolation: VM or microVM
Typed value: No.
Managed worktrees separate checkout directories/branches. The optional official Modal plugin adds image-based container sandboxes in the operator account; no selectable VM/microVM environment backend ships in this revision.
- Container: Official Modal sandbox plugin builds Dockerfile images and creates reusable sandbox machines from them.
- Vm: Complete shipped environment provider set has host checkouts/worktrees and Modal image-based sandboxes, without a VM/microVM backend.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/environment-modal-sandbox/README.md
Finding: The optional official Modal sandbox plugin builds a named Dockerfile image, creates a reusable sandbox machine in the user Modal account, enrolls its daemon and clones projects. Operator Modal credentials pay provider infrastructure; bb supplies no vendor-operated execution service. The shipped environment backend set has host checkout/worktree or Modal image-based sandbox, with no selectable VM/microVM backend.
Parallel runs: Yes cross-checked
Parallel runs
Typed value: Yes.
Managed worktrees allow concurrent threads without changing the main checkout.
- Concurrent sessions: Managed worktrees allow concurrent threads without changing the main checkout.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Managed worktrees create separate branches, copy selected untracked files and run setup/teardown hooks. Setup failures prevent thread start; this is file separation, not a claimed security sandbox.
Coordinator: Yes cross-checked
Coordinator
Typed value: Yes.
Manager threads coordinate child threads; provider-native subagents can be disabled to route delegation through bb. Task comments can send new instructions and resume an idle worker.
- Roles: Manager threads coordinate child threads; provider-native subagents can be disabled to route delegation through bb.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
-
bb plugins/provider-claude-code/PLUGIN_OVERVIEW.md
Finding: Claude provider runs the installed signed-in Claude CLI and reads plan/usage; permission modes and optional disabling of provider-native subagents are documented.
-
bb plugins/provider-codex/PLUGIN_OVERVIEW.md
Finding: Codex supports ChatGPT login or API key, shows ChatGPT usage limits and can disable provider-native subagents so delegation uses bb.
Native issue tracking: Yes cross-checked
Native issue tracking
Typed value: Yes.
Product-owned tasks support create/edit, board/list views, filters and task detail. Task comments can notify/resume the last responding worker. Task metadata includes labels, priorities and due dates. Delegate starts a worker from a preset and moves the task to in_progress. A task retains linked worker-thread cards and is distinct from its worker conversation.
- Create: Product-owned tasks support create/edit, board/list views, filters and task detail.
- Discuss: Task comments can notify/resume the last responding worker.
- Prioritize: Task metadata includes labels, priorities and due dates.
- Explicit dispatch: Delegate starts a worker from a preset and moves the task to in_progress.
- Task session separation: A task retains linked worker-thread cards and is distinct from its worker conversation.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
Unattended triggers: Yes cross-checked
Unattended triggers
Typed value: Yes.
Automations support repeating cron/timezone and one-shot schedules. Automation scripts execute on the bb server host; agent runs use installed providers.
- Schedules: Automations support repeating cron/timezone and one-shot schedules.
- Scheduler location: Automation scripts execute on the bb server host; agent runs use installed providers.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/automations/PLUGIN_OVERVIEW.md
Finding: Automations run cron/timezone or one-shot schedules, create/re-prompt threads or run stored scripts, and retain run output/history. Scheduler scripts execute on the server host.
Team roles: No
Team roles
Typed value: No.
The scoped core uses trusted-network shared-instance access and optional paired-owner bb connect. It has no named human workspace roles or enterprise SSO integration.
- Organization roles: Core/account access model has no human organization/workspace membership and role-permission ladder.
- Sso: No enterprise SAML/OIDC SSO configuration in scoped core or paired-owner connection authentication.
- Source inspection covers the scoped release only. Unanswered atomic questions remain unknown; no live scenario was run.
-
Finding: A bb server can be opened by multiple browser devices to send prompts and manage the same threads. Direct trusted-network API access has no per-human authentication; bb connect gates the tunnel to its paired account owner. Core thread actions accept shared thread input/queued follow-ups, without a human organization-role model or enterprise SSO integration.
SSO: No
SSO
Typed value: No.
The scoped core uses trusted-network shared-instance access and optional paired-owner bb connect. It has no named human workspace roles or enterprise SSO integration.
- Organization roles: Core/account access model has no human organization/workspace membership and role-permission ladder.
- Sso: No enterprise SAML/OIDC SSO configuration in scoped core or paired-owner connection authentication.
- Source inspection covers the scoped release only. Unanswered atomic questions remain unknown; no live scenario was run.
-
Finding: A bb server can be opened by multiple browser devices to send prompts and manage the same threads. Direct trusted-network API access has no per-human authentication; bb connect gates the tunnel to its paired account owner. Core thread actions accept shared thread input/queued follow-ups, without a human organization-role model or enterprise SSO integration.
Pull request workflow: Yes cross-checked
Pull request workflow
Typed value: Yes.
Managed worktree environments create separate branches for parallel work. GitHub PR status plugin exposes PR state, Mark ready and merge actions.
- Pull requests: GitHub PR status plugin exposes PR state, Mark ready and merge actions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Git change summaries, optional commit action, PR status/readiness/merge actions and disabled-by-default retention policies are configurable. Interrupted-thread resume is experimental and off by default.
Review interface: Yes cross-checked
Review interface
Typed value: Yes.
The Git Diff tab displays uncommitted, branch and per-commit diffs with Viewed marks and filters.
- Diffs: The Git Diff tab displays uncommitted, branch and per-commit diffs with Viewed marks and filters.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/git/PLUGIN_OVERVIEW.md
Finding: The Git plugin displays uncommitted, branch and per-commit diffs, Viewed marks and filters, with CLI operations against the same environment.
Product license: Apache-2.0: No. MIT cross-checked
Product license: Apache-2.0
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: MIT: Yes. MIT cross-checked
Product license: MIT
Typed value: Yes. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: GPL-3.0: No. MIT cross-checked
Product license: GPL-3.0
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: GPL-3.0-or-later: No. MIT cross-checked
Product license: GPL-3.0-or-later
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: AGPL-3.0: No. MIT cross-checked
Product license: AGPL-3.0
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: AGPL-3.0-or-later: No. MIT cross-checked
Product license: AGPL-3.0-or-later
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: Elastic License 2.0: No. MIT cross-checked
Product license: Elastic License 2.0
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: Custom terms: No. MIT cross-checked
Product license: Custom terms
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Product license: Proprietary: No. MIT cross-checked
Product license: Proprietary
Typed value: No. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Self-hosting: Yes cross-checked
Self-hosting
Typed value: Yes.
The central server, SQLite state and execution daemons can run on user-controlled machines. Direct trusted-network access is available; bb connect and inference providers are separate service routes.
- Self hosting: The central server, SQLite state and execution daemons can run on user-controlled machines.
- Vendor dependencies: Direct trusted-network access is available; bb connect and inference providers are separate service routes.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
Telemetry opt-out: Yes cross-checked
Telemetry opt-out
Typed value: Yes.
Optional storage-retention plugin sets archive/delete policies and defaults to disabled. Production app starts/thread/message counts and public plugin installs send anonymous telemetry by default; source dev modes do not. BB_TELEMETRY=false disables anonymous bb usage events.
- Telemetry defaults: Production app starts/thread/message counts and public plugin installs send anonymous telemetry by default; source dev modes do not.
- Telemetry controls: BB_TELEMETRY=false disables anonymous bb usage events.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
Free access: Yes cross-checked
Free access
Typed value: Yes.
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Ongoing free: The desktop app and server are free MIT downloads from public GitHub releases and npm.
- Self host cost: Server and execution hosts run on the user's machines with no bb fee shown; provider charges stay separate.
- Paid only: Not paid-only: the application installs free from public channels.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.
Starting price: Free: Yes. Free cross-checked
Starting price: Free
Typed value: Yes. Free.
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Ongoing free: The desktop app and server are free MIT downloads from public GitHub releases and npm.
- Self host cost: Server and execution hosts run on the user's machines with no bb fee shown; provider charges stay separate.
- Paid only: Not paid-only: the application installs free from public channels.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.
Starting price: Up to $20/month: No. Free cross-checked
Starting price: Up to $20/month
Typed value: No. Free.
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Ongoing free: The desktop app and server are free MIT downloads from public GitHub releases and npm.
- Self host cost: Server and execution hosts run on the user's machines with no bb fee shown; provider charges stay separate.
- Paid only: Not paid-only: the application installs free from public channels.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.
Starting price: $21–40/month: No. Free cross-checked
Starting price: $21–40/month
Typed value: No. Free.
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Ongoing free: The desktop app and server are free MIT downloads from public GitHub releases and npm.
- Self host cost: Server and execution hosts run on the user's machines with no bb fee shown; provider charges stay separate.
- Paid only: Not paid-only: the application installs free from public channels.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.
Starting price: $41–100/month: No. Free cross-checked
Starting price: $41–100/month
Typed value: No. Free.
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Ongoing free: The desktop app and server are free MIT downloads from public GitHub releases and npm.
- Self host cost: Server and execution hosts run on the user's machines with no bb fee shown; provider charges stay separate.
- Paid only: Not paid-only: the application installs free from public channels.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.
Starting price: Over $100/month: No. Free cross-checked
Starting price: Over $100/month
Typed value: No. Free.
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Ongoing free: The desktop app and server are free MIT downloads from public GitHub releases and npm.
- Self host cost: Server and execution hosts run on the user's machines with no bb fee shown; provider charges stay separate.
- Paid only: Not paid-only: the application installs free from public channels.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.
Starting price: Contact for pricing: No. Free cross-checked
Starting price: Contact for pricing
Typed value: No. Free.
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Ongoing free: The desktop app and server are free MIT downloads from public GitHub releases and npm.
- Self host cost: Server and execution hosts run on the user's machines with no bb fee shown; provider charges stay separate.
- Paid only: Not paid-only: the application installs free from public channels.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.
Pricing model: Free, no paid plan: Yes
Pricing model: Free, no paid plan
Typed value: Yes.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
Pricing model: Free plan: No
Pricing model: Free plan
Typed value: No.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: bb uses the provider CLI the user has already authenticated.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
Pricing model: Subscription: No
Pricing model: Subscription
Typed value: No.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: bb uses the provider CLI the user has already authenticated.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
Pricing model: Per seat: No
Pricing model: Per seat
Typed value: No.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: bb uses the provider CLI the user has already authenticated.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
Pricing model: Usage-based: No
Pricing model: Usage-based
Typed value: No.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: bb uses the provider CLI the user has already authenticated.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
Pricing model: Per machine or compute: No
Pricing model: Per machine or compute
Typed value: No.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: bb uses the provider CLI the user has already authenticated.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
Pricing model: Bring your own key: Yes
Pricing model: Bring your own key
Typed value: Yes.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- Model cost: The user authenticates Codex with ChatGPT or their OpenAI API key; provider charges remain separate from free bb.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
bb plugins/provider-codex/PLUGIN_OVERVIEW.md
Finding: The official Codex provider uses the operator codex login on the execution machine and explicitly accepts a ChatGPT account or OpenAI API key. bb remains a free MIT application; those model charges are paid through the provider account, not a bb plan.
Pricing model: One-time purchase: No
Pricing model: One-time purchase
Typed value: No.
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: bb uses the provider CLI the user has already authenticated.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
Keeps going with laptop closed: Yes
Keeps going with laptop closed
Typed value: Yes.
Default launcher starts server and local host daemon on the current machine. Enrolled execution machines run provider processes remotely under one bb server. A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work. Optional Modal machines use the operator own provider account; there is no bb-vendor-operated execution service in the scoped source.
- Byo remote: Enrolled execution machines run provider processes remotely under one bb server.
- Client disconnect: A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
Open-source license: Yes. MIT cross-checked
Open-source license
Typed value: Yes. MIT.
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Client license: bb application source is MIT licensed.
- Server license: The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
Any desktop app: Yes cross-checked
Any desktop app
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Desktop os: Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
Any phone client: Yes cross-checked
Any phone client
Typed value: Yes.
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Mobile apps: iOS TestFlight and Android APK clients connect to a bb server.
- Notifications: The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile, and a plain-HTTP direct profile cannot register.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: iOS TestFlight and Android APK clients connect to the server. Phone apps are control clients with documented surface limitations, not execution hosts.
All public findings
Findings may summarize a source. They are not presented as verbatim statements.
bb:ACC-01: Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile. cross-checked
bb:ACC-01
Desktop supports macOS arm64; Linux x64 and Windows x64 installers are alpha. iOS TestFlight and Android APK clients connect to a bb server. The server serves a local or remote browser app for live thread inspection and steering. Phones and tablets can use a remote browser without the optional local editor helper. The bundled bb CLI can be installed on PATH or run through npm. Threads can be followed live, steered and handed off through the app, CLI or API. The push-notifications plugin notifies registered phones and open web and desktop clients; phone push needs the server to reach exp.host through bb connect or an HTTPS profile.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
-
Finding: iOS TestFlight and Android APK clients connect to the server. Phone apps are control clients with documented surface limitations, not execution hosts.
-
Finding: Git change summaries, optional commit action, PR status/readiness/merge actions and disabled-by-default retention policies are configurable. Interrupted-thread resume is experimental and off by default.
bb:ACC-02: Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions. cross-checked
bb:ACC-02
Installed Claude Code, Codex, Pi, Cursor via ACP, OpenCode, Grok Build and other compatible ACP agents are documented. Provider can be chosen per task; this source pass did not run mixed-harness sessions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
bb ACP provider plugin overview
Finding: The ACP plugin launches ACP agents on the host as thread providers. Besides ready-made Cursor, opencode, omp, Grok Build and Hermes providers, any agent in the official ACP registry that ships through npx or uvx can be added in settings or with bb acp add, and other ACP agents can be added as custom JSON entries.
-
Official ACP registry (source used by bb's ACP plugin)
Finding: The official ACP registry that bb's ACP plugin reads lists Gemini CLI (id gemini) distributed through npx, so it can be added to bb from the registry. No Aider entry was found.
bb:ACC-03: Codex accepts ChatGPT login, while Claude uses the signed-in CLI and reports its plan. Codex also accepts an OpenAI API key. cross-checked
bb:ACC-03
Codex accepts ChatGPT login, while Claude uses the signed-in CLI and reports its plan. Codex also accepts an OpenAI API key.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/provider-claude-code/PLUGIN_OVERVIEW.md
Finding: Claude provider runs the installed signed-in Claude CLI and reads plan/usage; permission modes and optional disabling of provider-native subagents are documented.
-
bb plugins/provider-codex/PLUGIN_OVERVIEW.md
Finding: Codex supports ChatGPT login or API key, shows ChatGPT usage limits and can disable provider-native subagents so delegation uses bb.
bb:ACC-04: Install/download bb, authenticate at least one agent provider and open the served app. Settings Machines generates an installer to enroll a persistent execution daemon; a server can run on an always-on VM. cross-checked
bb:ACC-04
Install/download bb, authenticate at least one agent provider and open the served app. Settings Machines generates an installer to enroll a persistent execution daemon; a server can run on an always-on VM.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
bb:ACC-05: bb CLI exposes thread and project operations; plugin commands extend it. The central server exposes HTTP API and WebSocket updates. Direct API is unauthenticated behind the documented trusted-network boundary. The npm package exports a Node SDK including thread spawn. The MCP plugin exposes CLI search/read/run for remote or local clients. cross-checked
bb:ACC-05
bb CLI exposes thread and project operations; plugin commands extend it. The central server exposes HTTP API and WebSocket updates. Direct API is unauthenticated behind the documented trusted-network boundary. The npm package exports a Node SDK including thread spawn. The MCP plugin exposes CLI search/read/run for remote or local clients.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
-
bb plugins/mcp/PLUGIN_OVERVIEW.md
Finding: Local/remote MCP exposes bb CLI search/read/run; remote apps require account approval and can be revoked. Write access is all-or-nothing and can steer agents on controlled hosts.
bb:ACC-06: Built-in provider, task, automation, memory, capacity and MCP plugins extend bb. cross-checked
bb:ACC-06
Built-in provider, task, automation, memory, capacity and MCP plugins extend bb.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
-
bb plugins/automations/PLUGIN_OVERVIEW.md
Finding: Automations run cron/timezone or one-shot schedules, create/re-prompt threads or run stored scripts, and retain run output/history. Scheduler scripts execute on the server host.
-
bb plugins/memory/PLUGIN_OVERVIEW.md
Finding: Memory plugin stores global/project records with provenance, versions, edit/delete UI and version-checked CLI writes. Summary catalog is injected each turn.
-
bb plugins/mcp/PLUGIN_OVERVIEW.md
Finding: Local/remote MCP exposes bb CLI search/read/run; remote apps require account approval and can be revoked. Write access is all-or-nothing and can steer agents on controlled hosts.
bb:ACC-07: The repository documents an Electron desktop shell around the web app. cross-checked
bb:ACC-07
The repository documents an Electron desktop shell around the web app.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
bb:PLAN-01: Tasks plugin creates product-owned task records independent of running threads. Tasks use project-prefixed stable keys such as PROD-1. cross-checked
bb:PLAN-01
Tasks plugin creates product-owned task records independent of running threads. Tasks use project-prefixed stable keys such as PROD-1.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
bb:PLAN-02: Task descriptions, subtasks, comments and attachments supply durable worker briefs. Revision-controlled requirements are not established. cross-checked
bb:PLAN-02
Task descriptions, subtasks, comments and attachments supply durable worker briefs. Revision-controlled requirements are not established.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
bb:PLAN-03: The Tasks plugin supports subtasks and nested folders; arbitrary blocker enforcement remains unverified. cross-checked
bb:PLAN-03
The Tasks plugin supports subtasks and nested folders; arbitrary blocker enforcement remains unverified.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
bb:PLAN-04: Global/project memory has provenance, kinds, tags, versions and a summary catalog injected each turn. Memory records support review/edit UI and version-checked CLI update. Memory records can be deleted through the UI or bb memory forget. cross-checked
bb:PLAN-04
Global/project memory has provenance, kinds, tags, versions and a summary catalog injected each turn. Memory records support review/edit UI and version-checked CLI update. Memory records can be deleted through the UI or bb memory forget.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/memory/PLUGIN_OVERVIEW.md
Finding: Memory plugin stores global/project records with provenance, versions, edit/delete UI and version-checked CLI writes. Summary catalog is injected each turn.
bb:PLAN-06: Product-owned tasks support create/edit, board/list views, filters and task detail. Task comments can notify/resume the last responding worker. Task metadata includes labels, priorities and due dates. Delegate starts a worker from a preset and moves the task to in_progress. A task retains linked worker-thread cards and is distinct from its worker conversation. cross-checked
bb:PLAN-06
Product-owned tasks support create/edit, board/list views, filters and task detail. Task comments can notify/resume the last responding worker. Task metadata includes labels, priorities and due dates. Delegate starts a worker from a preset and moves the task to in_progress. A task retains linked worker-thread cards and is distinct from its worker conversation.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
bb:PLAN-07: bb tasks CLI reads and writes the same stable task keys shown to humans. Agent CLI supports task create, update, move, comment, delegate and attachments. cross-checked
bb:PLAN-07
bb tasks CLI reads and writes the same stable task keys shown to humans. Agent CLI supports task create, update, move, comment, delegate and attachments.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
bb:COORD-01: Managed worktrees allow concurrent threads without changing the main checkout. cross-checked
bb:COORD-01
Managed worktrees allow concurrent threads without changing the main checkout.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Managed worktrees create separate branches, copy selected untracked files and run setup/teardown hooks. Setup failures prevent thread start; this is file separation, not a claimed security sandbox.
bb:COORD-02: Task delegation selects provider/model/effort/permissions and attaches a worker thread. cross-checked
bb:COORD-02
Task delegation selects provider/model/effort/permissions and attaches a worker thread.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
bb:COORD-03: Manager threads coordinate child threads; provider-native subagents can be disabled to route delegation through bb. Task comments can send new instructions and resume an idle worker. cross-checked
bb:COORD-03
Manager threads coordinate child threads; provider-native subagents can be disabled to route delegation through bb. Task comments can send new instructions and resume an idle worker.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
-
bb plugins/tasks/PLUGIN_OVERVIEW.md
Finding: Tasks plugin has persistent IDs, list/board views, labels, priorities, subtasks, comments, attachments and explicit delegation into worker threads. bb tasks CLI reads/writes the same tracker and reports results.
-
bb plugins/provider-claude-code/PLUGIN_OVERVIEW.md
Finding: Claude provider runs the installed signed-in Claude CLI and reads plan/usage; permission modes and optional disabling of provider-native subagents are documented.
-
bb plugins/provider-codex/PLUGIN_OVERVIEW.md
Finding: Codex supports ChatGPT login or API key, shows ChatGPT usage limits and can disable provider-native subagents so delegation uses bb.
bb:COORD-04: Automations support repeating cron/timezone and one-shot schedules. Automation scripts execute on the bb server host; agent runs use installed providers. cross-checked
bb:COORD-04
Automations support repeating cron/timezone and one-shot schedules. Automation scripts execute on the bb server host; agent runs use installed providers.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/automations/PLUGIN_OVERVIEW.md
Finding: Automations run cron/timezone or one-shot schedules, create/re-prompt threads or run stored scripts, and retain run output/history. Scheduler scripts execute on the server host.
bb:COORD-05: Live threads can receive steering through browser, CLI and HTTP/SDK surfaces. Codex native questions become question cards whose answers are sent as messages. cross-checked
bb:COORD-05
Live threads can receive steering through browser, CLI and HTTP/SDK surfaces. Codex native questions become question cards whose answers are sent as messages.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
bb plugins/provider-codex/PLUGIN_OVERVIEW.md
Finding: Codex supports ChatGPT login or API key, shows ChatGPT usage limits and can disable provider-native subagents so delegation uses bb.
bb:RUN-01: Default launcher starts server and local host daemon on the current machine. Enrolled execution machines run provider processes remotely under one bb server. A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work. Optional Modal machines use the operator own provider account; there is no bb-vendor-operated execution service in the scoped source.
bb:RUN-01
Default launcher starts server and local host daemon on the current machine. Enrolled execution machines run provider processes remotely under one bb server. A browser/phone executes nothing itself; work lives on an always-on server and execution host. Sleeping the server can stop access and work. Optional Modal machines use the operator own provider account; there is no bb-vendor-operated execution service in the scoped source.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
-
bb plugins/environment-modal-sandbox/README.md
Finding: The optional official Modal sandbox plugin builds a named Dockerfile image, creates a reusable sandbox machine in the user Modal account, enrolls its daemon and clones projects. Operator Modal credentials pay provider infrastructure; bb supplies no vendor-operated execution service. The shipped environment backend set has host checkout/worktree or Modal image-based sandbox, with no selectable VM/microVM backend.
bb:RUN-02: Managed worktrees separate checkout directories/branches. The optional official Modal plugin adds image-based container sandboxes in the operator account; no selectable VM/microVM environment backend ships in this revision.
bb:RUN-02
Managed worktrees separate checkout directories/branches. The optional official Modal plugin adds image-based container sandboxes in the operator account; no selectable VM/microVM environment backend ships in this revision.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Managed worktrees create separate branches, copy selected untracked files and run setup/teardown hooks. Setup failures prevent thread start; this is file separation, not a claimed security sandbox.
-
bb Modal sandbox plugin README
Finding: An optional, experimental official plugin runs bb machines as Modal sandboxes in the user's own Modal account, built from an editable Dockerfile image; core enrolls a bb daemon in each sandbox and clones the project into it.
-
bb plugins/environment-modal-sandbox/README.md
Finding: The optional official Modal sandbox plugin builds a named Dockerfile image, creates a reusable sandbox machine in the user Modal account, enrolls its daemon and clones projects. Operator Modal credentials pay provider infrastructure; bb supplies no vendor-operated execution service. The shipped environment backend set has host checkout/worktree or Modal image-based sandbox, with no selectable VM/microVM backend.
bb:RUN-03: The launcher restarts crashed server/daemon child processes separately. This does not guarantee resuming the interrupted agent turn. Automatic interrupted-thread resumption is an opt-in experiment, off by default. cross-checked
bb:RUN-03
The launcher restarts crashed server/daemon child processes separately. This does not guarantee resuming the interrupted agent turn. Automatic interrupted-thread resumption is an opt-in experiment, off by default.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
Finding: Git change summaries, optional commit action, PR status/readiness/merge actions and disabled-by-default retention policies are configurable. Interrupted-thread resume is experimental and off by default.
-
Finding: Current source changelog describes stuck-turn timeouts, provider retry de-duplication and reconnect catch-up. Automatic interrupted-thread resumption remains opt-in.
bb:RUN-04: Per-worktree setup scripts install dependencies before dispatch; failure or 15-minute timeout blocks thread start. worktreeinclude copies selected untracked files without symlink traversal and reports skipped copies. cross-checked
bb:RUN-04
Per-worktree setup scripts install dependencies before dispatch; failure or 15-minute timeout blocks thread start. worktreeinclude copies selected untracked files without symlink traversal and reports skipped copies.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Managed worktrees create separate branches, copy selected untracked files and run setup/teardown hooks. Setup failures prevent thread start; this is file separation, not a claimed security sandbox.
bb:RUN-05: The changelog describes failing provider turns accepted but not started after five minutes. cross-checked
bb:RUN-05
The changelog describes failing provider turns accepted but not started after five minutes.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Current source changelog describes stuck-turn timeouts, provider retry de-duplication and reconnect catch-up. Automatic interrupted-thread resumption remains opt-in.
bb:RUN-06: Service logs rotate and SQLite thread events retain timeline progress. cross-checked
bb:RUN-06
Service logs rotate and SQLite thread events retain timeline progress.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The npm package bundles launcher, bb CLI and Node SDK. It uses installed Claude Code, Codex, Pi or ACP providers and existing logins; server and daemon have separate restart/logging behavior.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
bb:RUN-08: New idle-thread turns queue at host/global capacity and recheck after another thread settles. Concurrency plugin derives each host limit from processor count and supports fixed overrides. cross-checked
bb:RUN-08
New idle-thread turns queue at host/global capacity and recheck after another thread settles. Concurrency plugin derives each host limit from processor count and supports fixed overrides.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/concurrency-limit/PLUGIN_OVERVIEW.md
Finding: Automatic host capacity is one running thread per processor, with host/global overrides. New turns wait at capacity and reconsider when running threads settle.
bb:TEAM-02: The scoped core uses trusted-network shared-instance access and optional paired-owner bb connect. It has no named human workspace roles or enterprise SSO integration.
bb:TEAM-02
The scoped core uses trusted-network shared-instance access and optional paired-owner bb connect. It has no named human workspace roles or enterprise SSO integration.
- Source inspection covers the scoped release only. Unanswered atomic questions remain unknown; no live scenario was run.
-
Finding: A bb server can be opened by multiple browser devices to send prompts and manage the same threads. Direct trusted-network API access has no per-human authentication; bb connect gates the tunnel to its paired account owner. Core thread actions accept shared thread input/queued follow-ups, without a human organization-role model or enterprise SSO integration.
bb:TEAM-03: Approved remote MCP app grants can be revoked. Access is all-or-nothing, not per-tool write authority. Direct network API access is unauthenticated and requires a trusted network; bb connect adds an account gate. cross-checked
bb:TEAM-03
Approved remote MCP app grants can be revoked. Access is all-or-nothing, not per-tool write authority. Direct network API access is unauthenticated and requires a trusted network; bb connect adds an account gate.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
-
bb plugins/mcp/PLUGIN_OVERVIEW.md
Finding: Local/remote MCP exposes bb CLI search/read/run; remote apps require account approval and can be revoked. Write access is all-or-nothing and can steer agents on controlled hosts.
bb:TEAM-04: Threads retain an append-only stream of messages, tool calls and file changes in server state. cross-checked
bb:TEAM-04
Threads retain an append-only stream of messages, tool calls and file changes in server state.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
bb:SHIP-01: The Git Diff tab displays uncommitted, branch and per-commit diffs with Viewed marks and filters. cross-checked
bb:SHIP-01
The Git Diff tab displays uncommitted, branch and per-commit diffs with Viewed marks and filters.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
bb plugins/git/PLUGIN_OVERVIEW.md
Finding: The Git plugin displays uncommitted, branch and per-commit diffs, Viewed marks and filters, with CLI operations against the same environment.
bb:SHIP-03: Managed worktree environments create separate branches for parallel work. GitHub PR status plugin exposes PR state, Mark ready and merge actions. cross-checked
bb:SHIP-03
Managed worktree environments create separate branches for parallel work. GitHub PR status plugin exposes PR state, Mark ready and merge actions.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Managed worktrees create separate branches, copy selected untracked files and run setup/teardown hooks. Setup failures prevent thread start; this is file separation, not a claimed security sandbox.
-
Finding: Git change summaries, optional commit action, PR status/readiness/merge actions and disabled-by-default retention policies are configurable. Interrupted-thread resume is experimental and off by default.
bb:SHIP-04: PR merge actions are provided by the GitHub PR status plugin; exact-revision authority enforcement is unverified. cross-checked
bb:SHIP-04
PR merge actions are provided by the GitHub PR status plugin; exact-revision authority enforcement is unverified.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Git change summaries, optional commit action, PR status/readiness/merge actions and disabled-by-default retention policies are configurable. Interrupted-thread resume is experimental and off by default.
bb:SHIP-05: Thread Info retains the linked pull request even when the above-composer PR panel is hidden. cross-checked
bb:SHIP-05
Thread Info retains the linked pull request even when the above-composer PR panel is hidden.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Git change summaries, optional commit action, PR status/readiness/merge actions and disabled-by-default retention policies are configurable. Interrupted-thread resume is experimental and off by default.
bb:OWN-01: bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code. cross-checked
bb:OWN-01
bb application source is MIT licensed. The same repository MIT grant covers core server and host-daemon code.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: The application repository is MIT licensed.
bb:OWN-02: The central server, SQLite state and execution daemons can run on user-controlled machines. Direct trusted-network access is available; bb connect and inference providers are separate service routes. cross-checked
bb:OWN-02
The central server, SQLite state and execution daemons can run on user-controlled machines. Direct trusted-network access is available; bb connect and inference providers are separate service routes.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: Server owns SQLite state, HTTP/WebSocket API and host routing. Enrolled daemons execute work. Manager threads coordinate child threads, with append-only events and explicit lifecycle relationships.
-
Finding: An always-on desktop/server/VM runs bb while remote browsers/phones steer threads. Additional host daemons execute work. bb connect is account-gated; direct API access relies on a trusted network and is unauthenticated.
bb:OWN-03: Optional storage-retention plugin sets archive/delete policies and defaults to disabled. Production app starts/thread/message counts and public plugin installs send anonymous telemetry by default; source dev modes do not. BB_TELEMETRY=false disables anonymous bb usage events. cross-checked
bb:OWN-03
Optional storage-retention plugin sets archive/delete policies and defaults to disabled. Production app starts/thread/message counts and public plugin installs send anonymous telemetry by default; source dev modes do not. BB_TELEMETRY=false disables anonymous bb usage events.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
-
Finding: Git change summaries, optional commit action, PR status/readiness/merge actions and disabled-by-default retention policies are configurable. Interrupted-thread resume is experimental and off by default.
bb:OWN-04: bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
bb:OWN-04
bb is free and open source (MIT) with no paid plan found; agents run on the user's own provider subscriptions, billed by the provider, not bb. The official Codex provider also accepts the user OpenAI API key for separately billed model usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: bb uses the provider CLI the user has already authenticated.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
-
bb plugins/provider-codex/PLUGIN_OVERVIEW.md
Finding: The official Codex provider uses the operator codex login on the execution machine and explicitly accepts a ChatGPT account or OpenAI API key. bb remains a free MIT application; those model charges are paid through the provider account, not a bb plan.
bb:OWN-06: README calls the project active development and Linux/Windows desktop alpha. Current pinned-source changelog includes 0.49.0 with reliability and experimental restart changes. cross-checked
bb:OWN-06
README calls the project active development and Linux/Windows desktop alpha. Current pinned-source changelog includes 0.49.0 with reliability and experimental restart changes.
- Only the cited atomic findings are established. Remaining atomic requirements and the full rubric scenario remain unverified.
-
Finding: bb has desktop/browser/CLI/API surfaces, uses authenticated provider CLIs, supports macOS arm64 desktop and alpha Linux/Windows clients. Production telemetry is on and BB_TELEMETRY=false disables it. Electron shell is documented.
-
Finding: Current source changelog describes stuck-turn timeouts, provider retry de-duplication and reconnect catch-up. Automatic interrupted-thread resumption remains opt-in.
bb:OWN-08: bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider. cross-checked
bb:OWN-08
bb is MIT open-source software; the desktop app downloads free from GitHub releases and the server runs with npx, with no paid plan shown. Agent usage is billed by each provider.
- Optional bb cloud services need a bb account and have a daily usage limit; their terms were not checked. Agent free tiers were not checked. No installed-product test was run.
-
Finding: The site says bb is free, open source and local-first, and that each agent runs on the user's own provider plan, billed by the provider rather than bb.
-
Finding: The MIT license grants permission to use, copy, modify and distribute the software free of charge.
-
bb README: download and install
Finding: The desktop app downloads from public GitHub releases (desktop-v0.49.0 ships macOS dmg, Windows exe and Linux AppImage) and the server runs with npx bb-app; the README names no price or paid plan.