Hey Product Hunt, we built 0ctx because too many AI workflows lose context between sessions, branches, and tool switches.

0ctx keeps project memory attached to the repo itself, so the next session can continue the work instead of asking you to restate decisions, status, and intent.

What it does today:

repo-first setup with 0ctx enable

local runtime backed by SQLite
workstreams, sessions, checkpoints, and reviewed insights
retrieval for supported agent workflows through the local runtime

The current open-source path is intentionally local-first: install the CLI, enable one repo, and work normally.

We would especially love feedback on:

whether the positioning is clear
whether the setup path feels obvious
which integrations you would want next

If you try it, tell us where the setup or mental model is confusing. That is the highest-value feedback for us right now.

What does 0ctx actually do?

0ctx is persistent project memory for AI work. It keeps sessions, checkpoints, and decisions attached to the repo so the next tool or session can continue with the right context.

Why is local-first important here?

The product is built around repo-bound workflows, so local-first keeps the runtime close to the code, reduces moving parts, and makes the memory model easier to trust and reason about.

How do I start?

Install the CLI, go into the repo you work in, then run:

npm install -g @0ctx/cli
cd <repo>
0ctx enable

Is this another hosted AI IDE?

No. The core story is not a new chat surface. The core story is durable project memory attached to the repo through a local runtime.

Which integrations are supported today?

The current GA path in this repo is:

Claude Code
Factory / Droid

We are working towards making others tools part of it as well. Preview integrations are opt-in and are not the main open-source setup story.

What should I try first?

Try it in one active repo, run 0ctx enable, then use your normal agent workflow and inspect the captured workstreams, sessions, or checkpoints.