An AI pair that has actually read your code
An AI coding agent being built for teams: grounded in your repository, agentic where that saves time, gated by human review where it must be, with the controls an organisation needs to roll it out safely.
Team plans are set out on the Code plans page.
At a glance
- What
- An AI coding agent for teams
- Grounding
- Your repository: modules, idioms, the versions you run
- Changes
- Every change lands as a reviewable diff
- Team controls
- Seats, per-key limits, usage attribution by developer
- Runs as
- An independent agent for its domain, with its own harness
- Plans
- Four team plans, on the Code plans page
- Status
- In development, waitlist open
02 · Capabilities
Assistance, agency, and adult supervision
AI coding fails teams in two opposite ways: suggestions too generic to trust, or agents too autonomous to audit. Runix Code is designed against both.
Grounded in your repository
Designed so suggestions follow your conventions, not a tutorial's.
- Codebase awareness — completion and chat that reference your real modules and idioms
- Editor & terminal — the same assistant in the IDE and on the command line
- Dependency literacy — answers that respect the versions you actually run
Tasks, not just tab-completes
Agentic where it saves hours; reviewable at every step.
- Task execution — implement, refactor and test multi-file changes from one instruction
- Review gates — every change lands as a reviewable diff, never a silent write
- Test-first habits — changes validated against your test suite before they reach you
Rolled out, not smuggled in
The administrative layer most coding tools leave to the honor system.
- Team management — seats, per-key limits and usage attribution by developer
- Model routing — your policies decide which models serve which repositories
- Data terms in writing — code serves the request; it is not training data
03 · Fit
Who it is for
Built team-first: useful solo, designed for organisations.
Engineering teams
Adopt AI coding with usage you can attribute, policies you can set, and diffs you can review: instead of a dozen personal accounts and no visibility.
Platform & DevEx
Standardise how AI touches your codebase: one integration, model routing under your control, and conventions enforced rather than eroded.
Regulated environments
Data terms in writing, a boundary you configure for where prompts may travel, and an audit trail for what the agent did and why.
04 · Specifications
The details a review asks for
One table, the same shape on every product page, kept current with the pages it points to.
| Item | Value | Where it is documented |
|---|---|---|
| Assistance | Codebase-aware completion and chat, in the editor and on the command line | Rollout guide |
| Tasks | Multi-file implement, refactor and test from one instruction | This page |
| Review gates | Every change lands as a reviewable diff, never a silent write | This page |
| Team controls | Seats, per-key limits, usage attribution by developer | Rollout guide |
| Runs as | An independent agent for its domain, with its own harness; Runix Router is optional in front of its model calls | Runix Router |
| Plans | Four team plans | Code plans |
| Availability | In development; waitlist cohorts | Rollout guide |
05 · Works with
The layers on either side
Runix Code hands off to its neighbours in the stack and works on its own. The same block sits on every product page, with the neighbours changed.
Below · 04 Gateway
Runix Router
Optional in front of an agent's model calls: one endpoint for every model, with keys and quotas beneath it. The agent runs independently of it.
Explore Runix Router →Same layer · 05 Agents
Runix Comic
The other agent on the stack: an independent studio agent for comic dramas.
Explore Runix Comic →The whole stack
Five layers, seven products
How the agents sit on the gateway, models, data and storage beneath them.
See the stack →06 · How to start
How to start
Three steps, each with a person on the other end: Runix Code is not self-serve.
01Join the waitlist
Tell us the team size and the repositories. Cohorts are invited in order.
02Set up seats and policies
Seats, per-key limits and review gates configured with you, following the rollout guide.
03Roll out under review
Every change lands as a reviewable diff, so the team keeps its review habits from day one.
07 · Questions
Common questions
What is Runix Code?
An AI coding agent being built for teams: codebase-aware assistance, agentic task execution under review gates, and the administrative layer (seats, attribution, data terms) that makes org-wide rollout sane.
How is it different from a generic assistant?
Context and governance. Suggestions grounded in your repository rather than generic patterns; usage attributed per seat, models routed under your policies, prompts kept inside the boundary you configure.
When can I use it?
It is in development and not generally available. We onboard from the waitlist in small cohorts: tell us your stack and team size and we reply within one business day.
Is my code used for training?
No. Code processed by the product serves your request, is not used to train models and is not sold: the same posture as every Runix product. See Privacy.
Get your team in early
Tell us your stack, your team size and what you want an agent to take off your plate: we reply within one business day with where the product is and when you would get access.
Join the waitlist