Keep your AI agents
in sync everywhere.

One config for every machine and platform. Claude Code, Cursor, Codex, Gemini, OpenClaw - all from the same source.

$ npx @teamrc/cli init

Creates a .teamrc.yaml and walks you through setup.

Works with Claude Code Cursor Codex Gemini OpenClaw

How it works

Three steps to sync.

01

Define your agents once

A single .teamrc.yaml describes your team - agents, skills, and how they're wired together.

02

Each platform gets native files

Claude Code gets Markdown. Codex gets TOML. Gemini gets YAML frontmatter. All generated from the same source.

03

Sync across every machine

Run teamrc sync anywhere. The relay pushes changes and merges knowledge automatically.

Try it

Build your team.

Toggle agents and watch the config update live.

reviewer

Reviews PRs for correctness and style

architect

Plans system design and architecture

tester

Writes and maintains test suites

deployer

Manages deployments and infrastructure

docs-writer

Writes and updates documentation

.teamrc.yaml

team: my-project
agents:
- name: reviewer
description: Reviews PRs for correctness and style
- name: architect
description: Plans system design and architecture

Sync

Change once, update everywhere.

Edit one agent in your YAML. The change propagates to every platform.

.teamrc.yaml

agents:
- name: reviewer
- description: Reviews PRs for correctness
+ description: Reviews PRs for correctness and security
+ skills: [code-review, security-audit]
- name: architect
description: Plans system design
teamrc apply

.claude/agents/reviewer.md

---
name: reviewer
description: Reviews PRs for correctness and security
---
You are a code reviewer. Focus on
correctness, style consistency, and
security vulnerabilities.
## Skills
- code-review
- security-audit

.cursor/agents/reviewer.md

---
name: reviewer
description: Reviews PRs for correctness and security
tools:
- codebase_search
- read_file
---
You are a code reviewer. Focus on
correctness, style consistency, and
security vulnerabilities.

.codex/agents/reviewer.toml

[agent]
name = "reviewer"
description = "Reviews PRs for
correctness and security"
[agent.skills]
code-review = true
security-audit = true
[agent.tools]
sandbox = "full"

Features

What you get

Relay sync

Same team across different repos and projects. The relay keeps everything aligned across machines, no shared repo required.

Shared knowledge

An agent discovers a pattern and writes it down. Every agent on every machine knows it by the next sync. No more siloed context.

Format conversion

One YAML definition, five native output formats. No manual translation between platforms.

60+ agents, 50+ skills

Pre-built catalog across development, infrastructure, quality, and research. Use as-is or customize.

Build your team visually

Browse the agent and skill catalog, pick a template, and configure your team - no YAML required. The web wizard generates a .teamrc.yaml you can pull down with teamrc sync.

Open the wizard

Knowledge

Your agents learn once. Every machine remembers.

As agents work, they write findings to a shared knowledge file: build quirks, architectural decisions, environment-specific fixes. When you sync, knowledge from every machine merges automatically. The agent on your VM benefits from what the agent on your laptop discovered an hour ago.

With the daemon running, this happens in real time over WebSocket. No manual copying. No stale context.

Use cases

Useful when

Multiple machines

Laptop, VM, CI - all in sync

Mixed platforms

Cursor, Codex, Gemini in one team

Fast onboarding

Full agent setup in one command

Consistent behavior

Same agents everywhere they run

Try it now.

One command. No account required.

$ npx @teamrc/cli init

Creates a .teamrc.yaml and walks you through setup.

Open source · No account required for local use · Self-host or use the hosted relay