Code with Capuchin, Free. No card.
Capuchin serves 92.7% of production requests in MonkeysCode. Not a demo model, not a throttled trial tier. The model our paying users run all day, free for a month, with 50+ agent tools and no credit card.
Measured on production traffic.
“Why would I use your model instead of Claude?”
Fair question. Here is the honest answer.
For the hard 7%, you should use Claude. That is why it is in the plan.
For the other 93%, you are already using something like Capuchin and you are paying frontier prices for it. Autocomplete, applying an edit, renaming across files, answering a quick question about unfamiliar code. That work does not need a frontier model. It needs speed.
Capuchin handles it in under a second, unmetered. Claude, Gemini and ChatGPT stay in the plan for the problems that deserve them. That is not a compromise. It is the correct allocation, and it is why a $20 plan here stretches to 300+ frontier agent runs instead of a handful.
What 30 days actually gets you.
| Free plan | Your 30 days | |
|---|---|---|
| Capuchin | Not included | Unlimited |
| Claude, Gemini, ChatGPT | Your own keys | Included |
| 50+ agent tools | Yes | Yes |
| Background & scheduled agents | No | Yes |
| Session length | Standard | 100+ messages |
| Credit card | N/A | Not required |
What is MonkeysCode?
An agentic IDE that plans, edits, and verifies code across your entire codebase — with Capuchin built in free, and the frontier models included in your plan.
Capuchin, free to try.
Our fast-and-reasoning coding model ships with the IDE — free during the promo, unlimited on paid plans. It handles the everyday 80% of your work.
The frontier, in your plan.
Claude, Gemini, and ChatGPT included as a monthly allowance — one subscription instead of three. Lean local context keeps requests at ~$0.06 on average.
Yours, always.
Also runs 100% local or with your own key — free forever. Signed, replayable agent runs. No lock-in, ever.
Start coding with Capuchin AI — free.
Meet Capuchin.
A coding model built for shipping software — fast when it's easy, deliberate when it's hard. Plans, writes, and verifies across your whole codebase. The default the moment you open the editor.
Instant when it's easy. Deliberate when it's hard.
Flash mode — answers and edits land immediately for the everyday work — completions, small changes, quick questions.
Reason mode — for the hard stuff, Capuchin thinks step by step — planning multi-file changes, untangling bugs.
Automatic routing — MonkeysCode picks the mode by the task. Speed by default, depth on demand.
Two minds in one model
Built to use tools, not just talk about them.
Reliable tool use — calls your terminal, git, test runner, browser, and MCP servers — accurately.
Plans then executes — produces a plan you approve, then carries it out across files in parallel.
Self-checking — runs your tests and fixes what it broke before handing back.
Agent-native
Reads your codebase, not just the open file.
Massive context window — hundreds of thousands of tokens. Whole services and long histories fit at once.
Repo-aware — paired with MonkeysCode's on-device index, it reasons across files, services, and repos.
Multi-language — strong across TypeScript, Python, Go, Rust, PHP, Java, and more.
Sees the whole picture
Start coding with Capuchin AI — free.
Everything you need, on your terms.
Built from the ground up for developers who want power without lock-in.
Every Model, One Plan
Capuchin free + Claude, Gemini, ChatGPT in your plan — or your own model. Switch per task, never locked in.
Verifiable Runs
Signed event logs, test-gated commits, content-signed diffs. Every agent action is auditable.
100% Local Option
On-device indexing, local models, air-gapped mode. Your code never leaves unless you say so.
Every Token Counted
~$0.06 avg per frontier request, ~90% cached context, per-task budgets, hard cap, live dashboard.
Open Extensions
Open VSX by default, sideload any .vsix, self-host your own registry. No marketplace lock-in.
All 3 Platforms
Native on Windows, macOS, and Linux. Signed installers. Same engine powers editor, CLI, and CI.
50+ tools, not 50+ promises.
Filesystem, terminal, git, test runner, browser, search, subagents, MCP. Every one implemented, every one permission-gated, every one available during the trial.
An agent that can only suggest is a chatbot. Yours can edit across files, run your test suite, read the failure and fix it before handing the work back.
See all 50+ toolsTurn one month into two.
Every teammate who joins earns you both an extra week, up to four. Bring your team and you are close to two months of building free.
What happens on day 31.
Nothing you have to do anything about.
Capuchin stops. Nothing is charged, because we never took a card. You keep the IDE and carry on with your own keys exactly as before, or continue on a plan from $20.
We are not going to auto-enrol you into something you did not choose. That pattern exists because it works financially, and it poisons the well for everyone offering something genuinely free.
Keep Pro — $20/mo
Unlimited Capuchin plus a frontier allowance that goes further here — ~$0.06 per frontier request, good for 300+ agent runs a month.
See plansStay free — forever
The full IDE keeps working forever with local models or your own API key. Your editor never expires — no lock-in, ever.
How local & BYOK workStart building with Capuchin today.
No credit card required. No commitments. Full access to the AI-native IDE.
Try the model 92.7% of our users are already running.
Capuchin — free to try, the frontier models in one plan, 50+ agent tools.
Takes about 30 seconds · No card · Founding cohort is limited