Manyfold · your agent workspace
Many agents. One fold.
Any agent framework, or bring your own, all from one workspace.
Manyfold is an AI agent workspace for coding agents. Create and run Claude Code, Codex, Gemini CLI, OpenClaw, Hermes and other agents from one place — with chat, files, terminal, resumable sessions, skills and team channels.
Works with Claude Code · Codex · Gemini CLI · NarraNexus · Dify · Langflow · Hermes · Openclaw
From zero to a working agent in three steps.
No config sprawl. No bespoke wiring. Pick, connect, and start working.
- Pick a framework. Choose Claude Code, Codex, Gemini CLI, or any agent you already use.
- Connect a machine. Try one instantly, keep one running, or use your own machine. One click each.
- Start talking. Same composer for every agent. Switch frameworks mid-thread.
Run your workspace anywhere.
Spin up a workspace in seconds, keep one running, or wire in your own machines. Mix all three; Manyfold treats them the same.
- Try it instantly — A fresh workspace in seconds. Perfect for quick tasks and experiments — close it when you’re done, nothing to clean up.
- Keep it running — A workspace that stays awake between visits. Your files, tools, and sessions are right where you left them — ready for ongoing work.
- Use your own machines — Your laptop or your company’s servers. Everything stays on your side — nothing leaves the machines you already trust.
The workspace, and everything around it.
Automations, integrations, and connective tissue that make agents a daily habit, not a demo.
- Automate agent runs. Trigger any agent on a schedule, a webhook, a message, or a code event. Same workspace, hands-off.
- Reach you where you already chat. Mention an agent in Lark, Slack, Discord, Telegram, or Matrix. The reply lands in the same thread.
- One thread. Every agent. Switch between Claude Code, Codex, and Gemini in a single conversation. Context carries across.
- Bring your tools & MCPs. Plug in MCP servers, custom tools, your own model gateway. Every agent inherits the same toolbox.
Start free. Scale when ready.
Every plan runs the same platform. Pick the limits that fit how many agents you keep online.
- Free — Free · Great for getting started and seeing what Manyfold can do.
- Hobby — $9 / mo · For solo builders running a few agents in production.
- Plus — $19 / mo · For small teams collaborating across many agents.
- Pro — $49 / mo · For teams running production workloads at scale.
Prices in USD. Cancel anytime. Need a custom tier? Reach out and we will size it with you.
Frequently folded.
- Is this another wrapper around Claude / Codex?
- No. Manyfold hosts the actual agent runtimes — Claude Code, Codex, Gemini CLI — in isolated workspaces with attached files, terminal state, and resumable sessions. You bring your provider keys; we run the loop.
- What's the difference between a sandbox and a cloud computer?
- Sandboxes are durable but on-demand: ideal for repo tasks where the agent needs files, a terminal, and a session you can pick up later. Cloud computers stay online for cron jobs, connectors, and services that should never sleep.
- Do I need to bring my own provider keys?
- You can. Keys are stored per-workspace and routed per-agent — Claude Code on Anthropic, a Codex job on OpenAI, a Gemini CLI session on Google. Or use our metered access for any of them.
- Can I use my own framework agent?
- Yes. OpenClaw and Hermes are first-class, and any framework that runs in Linux + has a CLI entry point can be hosted as a custom agent definition.
- How does pricing work?
- A flat workspace fee covers the rail, sessions, and one stateful sandbox. Cloud computers and provider tokens are metered. No per-seat tax.
Start with your first agent
Free during preview. Bring as many agents and machines as you want. Talk to them all in one place starting today.