CursorClaude Code, Codex, and Cursor’s own agent work together on the plans you already pay for. Free and open source.
upload.ts. Starting a Codex review.pnpm test passes.upload.ts and its tests.Claude Code runs on your Claude plan, Codex on your ChatGPT plan, and Cursor’s agent on your Cursor plan. Each agent works the way its own team built it, and you pick the right one for each task.
When a Claude or Codex account hits its limit, bb moves the thread to your next one and it keeps going. bb itself is free.
Give one agent a big task. It splits the work across Claude Code, Codex, Cursor, and others running side by side, and they message each other as they go. You come back to finished work, not notes to pass between tools.
Need another? Add it with a plugin.
You can run any agent on your laptop, a desktop at home, or a cloud server, and manage them all from one bb. Check in from the bb mobile app or any browser while they keep working.
Remote machines, the mobile app, and browser access are all free.
bb comes with everything you need out of the box: worktrees, diffs, automations, a mobile app and more.
When you want more—or less—customize in Settings, browse the plugin marketplace, or ask an agent to build exactly what you need, immediately available wherever you use bb, including your phone.
CursorFeature by feature
$ Pro marks features that need a paid Cursor plan, from $20 a month.
| Agents | bb | Cursor |
|---|---|---|
| Agents | ||
| Claude Code and Codex agents | YesSide by side with Cursor’s agent | PartlyAs editor extensions |
| Claude and GPT models | YesBilled to your Claude and ChatGPT plans | Yes$ Pro plan onlyBilled to your Cursor usage |
| Agent-to-agent handoff | YesAny provider, each in its own thread | PartlySubagents inside Cursor’s agent |
| Switch accounts at usage limits | YesAutomatic with Account Pooler | NoOn-demand billing instead |
| Price and license | ||
| Pricing | $0You pay agent providers directly | $20 / moTeams at $40 per user / month |
| Open-source license | YesMIT | NoClosed source |
| Away from your desk | ||
| Run agents on other machines you own | YesAny computer, free | Yes$ Pro plan onlySelf-hosted cloud agents |
| Cloud agents | YesModal plugin, experimental | Yes$ Pro plan onlyHosted VMs at API rates |
| Mobile app | YesiOS beta, Android alpha | YesiOS beta, web on Android |
| Scheduled automations | YesOn your own machines | Yes$ Pro plan onlyRun as cloud agents |
| Workspace | ||
| Parallel agents in Git worktrees | YesSetup runs for you | YesAgents window |
| Diff review and merge | Yes | Yes |
| Plugin marketplace | YesPanels, commands, and agents | YesRules, skills, and MCP servers |
| Code editor | PartlyOpens your worktree in Cursor | YesFull editor with Tab |
Cursor is a code editor with its own agent, plus cloud agents on its paid plans. bb isn’t an editor: it’s a free, open-source app for running coding agents. Claude Code, Codex, Cursor’s agent, and others run side by side on your own machines, each on its own plan, and they can start and message each other. You can keep editing in Cursor while bb runs the agents.
Yes, for running coding agents: bb. It’s free for any team size and MIT-licensed, so you can use and change it for anything, including at work. You pay only for the agent plans you already have. Download bb.
Yes. Keep editing in Cursor, and open any thread’s worktree in it with one click. You can also pick Cursor as the agent for any bb thread: bb runs Cursor’s agent CLI with your own sign-in and plan, so your rules and MCP servers come along.
Ask bb to do it. Cursor’s local worktrees are plain Git on your machine, and its cloud agents push their work to branches, so a bb agent can add your repo and open each unfinished worktree or branch as a thread. Your AGENTS.md, CLAUDE.md, skills, and MCP servers come along, and Cursor keeps working while you try bb.
.cursor/worktrees.json move to .bb-env-setup.sh, and untracked files like .env go in .worktreeinclude.Yes. bb runs the agents you already use, signed in the way you already pay for them: a Claude Pro or Max plan, a ChatGPT plan for Codex, a Cursor plan, API keys, or any other provider’s plan. Your existing setup comes along too (CLAUDE.md, skills, MCP servers, and agent settings).
Yes, with Account Pooler, an experimental plugin built into bb. Turn it on and add the Claude Code and Codex accounts you own. When one account hits its limit, bb moves the thread’s requests to the next one, so it keeps running.
Yes. Ask Claude Code to “start a bb Codex thread to review this branch, then fix what it finds.” It starts Codex in the same worktree, waits, and applies the fixes. Codex shows up as its own thread, so you can read the exact prompt Claude sent, watch it work, and message it mid-run. Any other pair works the same way, like Cursor and OpenCode.
Yes, and there’s no coordinator to set up. Tell your agents to work together and they coordinate on their own: any agent can message another thread, whatever the provider, agents answer each other’s messages, and an agent hears back automatically when a thread it started finishes or fails. The receiving thread shows each message and who sent it, and you can step in and message any of them yourself.
Claude Code, Codex, Cursor, and OpenCode, plus Pi, Grok Build, omp, and Hermes Agent. In Settings, add any other agent that supports the Agent Client Protocol (ACP), an open standard for connecting coding agents to apps, like Gemini CLI or Devin. You can also add one with a plugin. Each thread can use a different agent.
Yes. Pick a permission mode for each thread. Accept Edits applies changes inside the project and asks before anything else. Approve for me reviews requests automatically and sends high-risk ones to you. Full Access skips approvals and can run anything on your machine. Allow or deny approvals from your computer or your phone.
Yes. Every bb thread runs in the background on the machine you pick: your laptop, an always-on desktop, or a server you own. You can also start threads in on-demand cloud sandboxes in your own Modal account with the built-in Modal Sandbox plugin (experimental). Either way, you pay nothing extra to bb.
Yes. Give each thread its own Git worktree so agents don’t overwrite each other’s changes. List your .env files and setup commands once, and bb prepares every new worktree. By default, bb runs one thread per processor core and starts the rest as others finish.
Yes. Commit a .bb-env-setup.sh at your repo root and bb runs it in every new worktree. List untracked files like .env in .worktreeinclude and bb copies them in first, and .bb-env-teardown.sh cleans up when a worktree goes away.
Yes, without leaving the thread. Select lines in its diff and choose Add to chat to send feedback. Once the agent opens a pull request, the thread shows its checks and a Merge button. The GitHub plugin adds Review with agent to any PR.
bb opens any thread’s worktree in Cursor, VS Code, or another editor you have installed. For quick edits without leaving bb, turn on the File Editor plugin.
Yes, for free. Use the bb mobile app, a public beta on iPhone through TestFlight (Apple’s beta testing app) and in alpha on Android, or open bb in any browser through bb Connect, bb’s free remote access.
Yes, free. Automations start an agent thread or run a script on a repeating schedule, once at a set time, or after a delay. Pick the agent, model, and permission mode, and give each run its own worktree if you like.
Not on that laptop. Agents run on your machine, and closing the lid puts it to sleep. On a Mac, switch on Keep Awake, a built-in bb plugin that stops idle sleep while bb runs. For long runs, add an always-on desktop or server to bb, or run threads in the cloud with a cloud plugin like Modal Sandbox, which is built in (experimental) and starts each thread in an on-demand sandbox in your own Modal account.
Yes. bb works out of the box, with worktrees, diff review, automations, and the mobile app ready from your first thread. When you want more, install plugins from the marketplace or ask an agent to build one. Plugins can add panels, commands, and new agents, and they work in the mobile app too. Everything in the app is also in the bb CLI and HTTP API, so scripts and other agents can start, message, and manage threads.
Yes, for one person or a whole team. The mobile app, remote machines, automations, and plugins are all included, while Cursor puts cloud agents and frontier models on its paid plans. You pay only for the agent plans or API keys you already use, from any provider. The code is on GitHub under the MIT license.
Yes. Download the app for Apple Silicon Macs, Windows, or Linux (alpha), or run npx bb-app@latest on an Intel Mac.
Yes: bb runs on your machines, so your code goes only where your agents send it, the AI providers you chose. Anonymous usage stats exclude code and prompts and can be turned off. The bb Connect relay doesn’t store traffic, but isn’t end-to-end encrypted.
Yes, free at any team size. Each person runs bb on their own machines with their own agent subscriptions and shares work through Git and pull requests as usual. bb doesn’t offer team plans, SSO, or a support SLA.
Free and open source. Bring the AI plans you already pay for.