Runix Code is an AI coding agent being built for teams. It is in development and not generally available; this page describes how adoption is designed to work, so you can decide whether to join the waitlist. Everything here is the plan being built with early cohorts — not a shipped product.
1. Joining a cohort
We onboard from the waitlist in small cohorts. When your team's turn comes we set up together: your stack, your repositories, and which policies you want enforced from day one. Cohort feedback decides what lands next, which is the honest reason to join early.
2. Seats and policies
- Seats — access is per developer, with usage attributed per seat so you can see where the value lands.
- Model routing — policies decide which models serve which repositories; sensitive code can be held to a stricter boundary than experiments.
- Data boundary — where prompts may travel is configured, and code is never training data. The posture is written down in Privacy.
3. The working loop
Day to day, Runix Code is designed to work in two modes. Assist: completion and chat in the editor and terminal, grounded in your repository rather than generic patterns. Agent: hand it a task — implement, refactor, add tests — and it works across files, validating against your test suite, and lands the result as a reviewable diff. Nothing merges without a human decision; the review gate is the product's spine, not an option.
4. What to tell us
When you write in, include your language stack, repository scale, team size, and the first workflow you would hand to an agent. We reply within one business day with where the product is and when your cohort would start.