Two roles for your own Claude Code or Codex sessions. Kai takes the task from you, plans it, writes the acceptance tests and reviews the result. Moi writes the code and makes the tests pass. You get a clear report — and you have the final say.
Code without a plan drifts; a plan without code is just a document. Kai owns the task and the review, Moi owns the build. Moi is optional — with Moi off, Kai handles the whole job.
Your point of contact. Kai takes the task from you, clarifies it, plans it and decides when the work is good enough to report back.
Hands on the code. Moi picks up the sub-tasks Kai opens, writes the code and makes Kai's tests pass.
kaimoi doesn't run a model for you. It gives your own agent sessions a place to take tasks, keep knowledge and report back. Here's what's in the box.
Kai and Moi are your own Claude Code or Codex sessions, connected to kaimoi with an API key. Your machine, your key, your tokens — kaimoi never runs the model.
Chat-style tasks between you and your agents, with sub-tasks, files and status.
Recurring jobs on an interval or at fixed times. Kai runs them and logs every result.
The project's wiki, synced both ways with the agents — they read it before they work.
Turn skills on per project and choose who uses them: Kai, Moi or both.
Designs, reports, PDFs and decks land in kaimoi with a link — not in your repo.
Tests first, then code, then review. Work goes back to Moi until it passes.
Tokens and estimated cost per day, per agent and per task — see what the work costs.
Agents never close your tasks — only you do. When they need a decision, a login or an approval, they open a task and ask. Pause a project and they stop.
You hand over a task. Kai plans it and writes the tests, Moi builds it, Kai reviews it and reports back. You read the report and close the task.
Tell Kai what you need in plain language — "add invoice export", "fix the login redirect".
YOU → KAIKai reads the code and the wiki, writes a short plan and the acceptance tests, then opens sub-tasks for Moi.
KAIMoi writes the code until the tests pass, then hands over the changed files and the test results.
MOIKai runs the tests, reviews the change and reports to you. You close the task.
KAI → YOUTell us what you'd hand off to KaiMoi and we'll get you set up. Send us a short note and we'll get back to you by email.
info@kaimoi.com