CLI Reference

Generates an ed25519 keypair (stored in ~/.teamrc/key), writes .teamrc.yaml to your project, and applies config to detected platforms. In interactive mode, you'll be asked whether to connect to teamrc.ai for cross-machine sync. Use --local to skip the prompt and create a local-only team. Use --relay to skip the prompt and connect immediately. The interactive wizard lets you pick a catalog template or build a custom team. If a .teamrc.yaml already exists without a team ID (e.g. from clone), init adopts it and no template selection is needed. When using --global, init also checks the current directory for a .teamrc.yaml to offer as the global team definition.

Connects this machine to an existing team using an invite code. Fetches the team definition from the relay, writes .teamrc.yaml, and applies config to all detected platforms. Each machine gets its own keypair. No shared credentials.

Like join, but doesn't register this machine with the relay. You get a local copy of the team definition. Cloned teams support read-only pull (to fetch updates from the original) but not push or sync. To fork a cloned team into your own, run teamrc init, which will adopt the existing definition, create a new team on the relay, and give you full ownership.

The primary command. Pushes local changes (including knowledge updates) to the relay, pulls the latest team definition, and regenerates all platform config files. Run this after making changes in the web UI, or after another machine has pushed updates.

One-way: fetches the latest team definition from the relay and writes it to local platform config files. Does not push any local changes. Use when you only want to receive updates.

One-way: reads your .teamrc.yaml and knowledge file, and sends them to the relay. Does not pull or apply anything. Use after manually editing .teamrc.yaml. If the team is local-only (not yet connected to a relay), push will register it on teamrc.ai first and then push. This is how you connect a local team.

Offline operation. Does not talk to the relay. Reads .teamrc.yaml and writes platform config files. Useful after editing the YAML directly, or to regenerate files for a newly installed platform.

Fetches the team definition from the relay and writes it to .teamrc.yaml without applying to any platforms. Use when you want to inspect or backup the remote state.

Shows what's different between your local .teamrc.yaml and the relay. Lists added, removed, and modified members. Helpful before pushing changes or debugging sync issues.

Creates a temporary invite code under the hood and builds a dashboard URL from it, so the link grants join access to the team. Opens the URL in your browser. Use this to manage the team in the web UI. Use invite instead when you want to share access with another machine or teammate.

Creates a time-limited invite code (trc_inv_...) that can be used with join. Share it with collaborators, use it on a VM, or connect another project to the same team. For opening your own browser session, use dashboard instead.

Displays your token, relay server, current team info (from .teamrc.yaml), and platform bindings. Quick way to see what's configured.

Reads existing agent configuration from a platform's native files and converts them into a .teamrc.yaml. Supported platforms: claude-code, cursor, codex, gemini, openclaw. Useful when migrating agents you've already set up manually.

Adds a new AI agent to the team. Without flags, launches an interactive wizard that lets you pick from the catalog or create a custom member. The new member is added to .teamrc.yaml and pushed to the relay.

Makes the current team publicly visible. Anyone with the share link can view the team composition and clone it. Requires a linked account (run teamrc login first) and team ownership. Run with --off to make the team private again.

Displays all available team templates from the catalog. Each template includes a name, description, and the agents and skills it contains. Use a template ID with teamrc init --team to start from a template.

Displays all available agent templates from the catalog. Each template includes a name, role, category, and pre-configured instructions and skills. Use a template ID with teamrc add-member --template to add a catalog agent.

Optional. Opens a browser-based device authorization flow (similar to gh auth login). Once confirmed, this machine is linked to your account for recovery and management via the web dashboard. You can use teamrc without logging in. It's only needed for the dashboard.

Displays your token, linked account email (if any), machine name, relay server, and current team. Like status but focused on identity.

Validates your entire setup: checks the keypair exists, config is valid, relay is reachable, and platforms are detected. Run this when something isn't working.

Associates your account as the owner of a team created via the CLI. The ownership token (trc_ocs_...) is displayed once during teamrc init. Save it like a recovery code. Requires a linked account (run teamrc login first). Once claimed, the token is invalidated and cannot be reused.

Removes all teamrc-generated files from all platforms, deletes .teamrc.yaml, and cleans up config. This is destructive and fully uninstalls teamrc from the current project.

Runs a background process that keeps the team knowledge file in sync across machines in real-time via WebSocket. Watches the local knowledge file for changes and pushes updates to the relay. Receives remote changes and merges them locally. Does not sync team config (members, skills). Use teamrc sync for that.