Switch between Claude Code and Codex. It already knows.
Team0 gives both the same up-to-date understanding of your work. Decide something in one, and the other starts from it.
The switch, today.
Most people who build with agents use both. Neither knows what happened in the other.
- You switch mid-task
- A limit, a price change, or the other one is just better at this. The new session starts from zero.
- You explain it all again
- The project, the customer, what you decided yesterday and why. Every new session, in every agent.
- Your notes files drift
- CLAUDE.md says one thing, AGENTS.md another. Keeping them current becomes a job of its own.
With Team0.
One understanding underneath both agents, kept current without you doing anything.
- Both read before they answer
- With the Team0 plugin, Claude Code and Codex read what Team0 knows before every reply. You don’t have to ask.
- Both save when a turn ends
- Decide something in Claude Code, and Codex starts from it in its next session. The same works the other way.
- The latest version wins
- When a plan changes, Team0 keeps the new one and moves the old one to history, so no agent acts on what you already changed.
- More than the repo
- Connect Google and your other sources, and both agents also know your customers, your people and what changed this week.
Set up once in each.
Team0 walks you through it and gives you the exact prompt to paste. After that, every session is automatic.
- Claude Code
- Paste the prompt Team0 gives you into Claude Code on your computer. Open a new local session, approve the plugin review, and click Connect Claude Code on the page that opens. Cloud sessions aren’t supported yet.
- Codex
- Paste the prompt Team0 gives you into Codex. Allow Team0’s three hooks (the hook icon in Codex desktop, or /hooks in Codex CLI), send a message, and click Connect Codex on the page that opens.
Questions developers ask.
- Does Team0 read my code or files?
- No. It doesn’t open your files or watch the agent’s tools. After each turn it receives your message and the agent’s final reply, so anything you paste into a message comes with it.
- What does it keep?
- Team0 pulls out what matters and deletes the conversation itself after 7 days. What it learned stays, with the line each piece came from. What an agent says is kept as “the agent said”, never treated as fact on its own.
- My agents already share a notes file. Why do I need this?
- It works at first. Then the file grows, key details get summarised away, and you end up maintaining it. Team0 keeps one understanding current from the work itself, so there is nothing to tend.
- Can I stop one agent from saving?
- Yes, per agent. One switch stops an agent from saving while it keeps reading. Another cuts it off completely. Your other agents are not affected.
- Can I see what an agent read?
- Yes. Every read keeps the exact request and what Team0 returned, and you can see each read an agent made over the last 30 days.
- Does it work in Claude Code cloud sessions?
- Not yet. The Claude Code plugin works in Claude Code on your computer: a terminal, an IDE, or Claude Desktop with Local selected.
- What about my other agents?
- OpenClaw and Hermes read and save on their own too. ChatGPT, Claude, Claude Cowork, Town, Lindy, monday.com and Base44 read when you ask them to. All of them read the same understanding.
Your agents will change. Your understanding shouldn’t.
Invite-only private beta.