Quickstart

0ctx is repo-first. Enable a repository, use your AI tools normally, and let context persist automatically.

The normal path

cd <repo>
0ctx enable
0ctx status

That's it. Once enabled, 0ctx runs in the background.

What `0ctx enable` does

Resolves the repo root
Creates or binds a workspace for that repo
Starts or verifies the local daemon
Installs supported GA capture integrations
Bootstraps supported GA retrieval integrations
Reports repo readiness and local retention defaults

Daily use

For supported GA agents, the intended flow is:

Run 0ctx enable once in the repo
Use the agent normally in that repo
Let 0ctx capture sessions and inject retrieval context automatically

Zero configuration

You don't need to think about contextId, MCP setup, hook internals, or transcript plumbing. The normal product path is 0ctx enable and then work.

Useful commands

# Repo readiness
0ctx status

# Manage local capture hooks
0ctx hook status

# Inspect workstreams
0ctx workstreams --repo-root .

# Inspect sessions
0ctx sessions --repo-root .

# Inspect checkpoints
0ctx checkpoints --repo-root .

# Advanced repair
0ctx doctor --json
0ctx repair

GA and preview integrations

GA path (automatic after enable):

Claude Code
Factory / Droid
Antigravity

Preview path (explicit opt-in only):

Codex
Cursor
Windsurf

Preview integrations stay outside the normal setup path. Only use them when you explicitly opt in.