The SwarmPane workspace
Watch your agents do the work.
SwarmPane brings Claude Code, Codex, and the other coding CLIs you use into one desktop workspace, with real terminals, project context, visual workflows, and a conductor that checks every result.
Get started · Try the live demo
macOS · Windows and Linux in testing · Bring your own provider accounts
Works with: Claude Code, Codex, Gemini CLI, Cursor Agent, GitHub Copilot CLI, OpenCode, Grok Build, Kimi CLI, Aider, Droid.
- 10 coding CLIs supported
- 4 workspace modes
- 16 steps per workflow
Four modes
One app. Four ways to work.
Switch modes as the job changes. Terminals keep running, drafts stay put, and every mode works in the same project.
Agent: Agents with a job and a memory.
Create teammates with their own instructions, skills, and long-term memory. Give one a task in your project and follow every turn, with your approval before it edits files.
- Instructions, skills & memory
- Claude Code or Codex, per turn
- Guided approval before edits
Code: Real terminals, side by side.
Run Claude Code, Codex, Cursor, Gemini CLI, and more in real terminals. Arrange them in Solo, Pair, Workbench, or Swarm layouts with files, Git, and a browser preview beside them.
- Persistent terminal sessions
- Files, diffs & commits
- Browser preview & project search
Chat: Think it through first.
A focused conversation for planning, drafting, and questions. Attach files and images, switch providers as you go, and pick up where you left off.
- Saved conversations
- Files & images as context
- Readable code in every reply
Workflow: Chain agents into a process.
Connect up to sixteen steps on a canvas. Each result moves to the next agent, approval steps pause the run for you, and every run is saved.
- Visual step editor
- Output passed between agents
- Approval steps & run history
Jev Conductor
Hand over the goal. Get a checked result.
Describe the outcome and how to check it. Jev picks one of your agents, runs the attempt in your project, and verifies it. When a check fails, the next attempt starts with the evidence.
- Attempt: Jev routes the task to one of your enabled agents.
- Verify: Your tests or checks decide whether it worked.
- Hand off or finish: A failed check moves to the next agent with context. You review the result.
Uses your enabled agents. Routing requires a SwarmPane account; provider usage is billed by your provider.
Everything around the agent
- Files and Git: Inspect and edit project files, review diffs, and commit without leaving the project.
- Approvals: Choose how much autonomy each agent has, and review changes to your work before they happen.
- MCP connections: Connect local MCP servers and check what their tools can reach before an agent uses them.
- Skills and memory: Give agents reusable instructions and let them keep your project’s rules and preferences.
- Tasks and routines: Track outcomes on a task board and schedule agent work that runs while the app is open.
- OverVoice (macOS): Dictate into a conversation or terminal pane, then edit the draft before you send it.
Local first
Your machine. Your accounts. Your call.
- Runs on your computer: Projects, terminals, and history stay on your Mac. The local core only accepts connections from this machine.
- Bring your own accounts: Agents use the Claude Code, Codex, and other CLI sign-ins you already have. Your provider bills its own usage.
- Scoped access: You choose the folders an agent can reach, and approve the actions that change your work.
Bring your agents. Build here.
Create your account, install the desktop app, and connect the coding tools you already use.
Get started · See pricing · Platform availability
SwarmPane Pro $29/month · Planned launch price. Checkout is not open yet. Use the AI accounts you already have.
Packaged desktop access requires SwarmPane Pro. Provider subscriptions are separate.