Pricing — Renitor is free during the open launch
Free — everything included, during the open launch.
No key, no trial clock, no card, no account. The whole point is that you're already paying for enough subscriptions.
- Repo-aware handoffs · 11 agent dialects
- Limit detection in your terminal
- Auto-capture of failing commands
- Two-layer secret redaction
- Multi-model chat — a failover chain that ends in free models, in one thread
- The multi-model toolkit — Team chat (boss & worker duo), Consensus, Debate, Group, second opinions, Improve, @mentions, smart routing, Guardian
- Session history & voice input — named, renameable, saved conversations
- Round-trip handback — native re-injection into Claude Code
- Goal auto-inference from agent session logs
- Agent safety ladder — Plan/Manual/Auto-edit/Auto, in-thread approvals, checkpoints with one-click revert
- PR / branch & focused handoffs
- Honest future: if a paid tier returns after the launch period, the build you installed keeps working.
More Claude capacity means a $100/month Max plan or extra usage charges. This is free — it works on the tools you already pay for.
Rolling it out to a team? Say hi — we'd love the feedback →
Questions we'd ask too.
Does this bypass Claude's or OpenAI's rate limits?
No — nothing can, and anything claiming to should worry you. It makes the limit painless instead: your other subscription takes over in seconds with the task carried over — the goal, decisions, changes and remaining work — and hands back after the reset.
Is this against the providers' terms of service?
It's designed to be ToS-conscious, not to promise the impossible: it never touches your auth tokens or wraps their APIs — it just invokes the official CLI you're logged into, the same pattern those CLIs use themselves. That said, providers control headless subscription use and can change the rules (Anthropic has already built, then paused, a change that meters headless usage), so subscription mode is best-effort — BYOK is the always-available fallback.
What data leaves my machine?
Nothing goes to us — no backend, no telemetry, no analytics, ever. The handoff is fully local (session traces come from log files your agents already keep on disk). The chat talks only to the provider you choose — your subscription CLI or your own API key — never through our servers.
Is this an enterprise tool?
Deliberately not. No sales calls, no seats, no SSO paywall — one person, one license, all their machines. If you need procurement and DPAs, the big platforms will take good care of you.
I pay once — who fixes it when Claude Code updates break something?
That's exactly what the included 12 months of updates are for: agent CLIs ship monthly, and adapter fixes are the core of what we maintain. The design also degrades gracefully — handback is strictly additive (it never edits your existing sessions), so the worst case of a format change is falling back to a paste-ready markdown brief until the fix ships, never a broken editor. And because everything is local-first, the version you have keeps working forever.
What happens when an agent changes its session format?
Handback injection is strictly additive — it writes a new session file and never edits yours. Worst case the feature degrades to a markdown brief until an adapter update ships.