---
title: "Caveman"
description: "Caveman Code. The most efficient coding agent. Same models, same work, a smaller bill. Every agent you use in one app, run by a Chief, from one account in the cloud. From the team that made agents talk like cavemen. Backed by Y Combinator."
canonical: https://caveman.so
last-updated: 2026-11-12
---

# Caveman

Caveman Code. The most efficient coding agent.

Same models, same work, a smaller bill. Every agent you use in one app, run by
a Chief, from one account in the cloud. From the team that made agents talk
like cavemen. Backed by Y Combinator.

- **[Caveman Code](https://caveman.so/code), the coding agent.** Every agent
  you use in one window, with a Chief to run them. Free on the subscriptions
  and keys you already have.
- **[Benchmarks](https://caveman.so/code#benchmarks): same model, smaller
  bill.** 21% less than Codex on Terminal-Bench 4, 76% less than Claude Code
  on work across connected apps.
- **[Caveman Cloud](https://caveman.so/cloud), where it runs.** See and
  control every agent. Coming soon: deploy the software factories you build
  with Caveman Code.

## Caveman Code

### Our agent runs on the most efficient harness

Same model, same work, 21% to 76% less.

A harness is everything around the model: the prompt, the tools and the loop
that runs them. Caveman, the agent built into Caveman Code, runs on ours. We
ran it against Claude Code and Codex, each on its own lab's model: the model is
the same, so the difference in the bill is the harness.

| Against | Model | Work | Their cost | Caveman Code | Change | Tokens | Result |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Codex | GPT-6.1 Sol | Terminal-Bench 4: 48 coding and system tasks in a terminal, 4 runs each, at high reasoning | $1.52 a solved task | $1.21 a solved task | 21% less | 31% fewer (3.2M to 4.7M a solved task) | 90 of 192 runs solved, both |
| Claude Code | Claude Opus 5.5 | Three working days: three day-long sessions, each one conversation of 15 to 17 requests across four connected apps | $7.65 a session | $1.83 a session | 76% less | 83% fewer (3.7M to 21.6M a session) | 48 of 48 steps passed, to Claude Code's 46 |

Same model and the same work on both sides, no router. Cost at list price.
Tokens are every token the model processed, input and output, counted over the
same runs. Our own runs on the harness's benchmark build, November 2026. The runs, the
versions and the notes on method are on
[Caveman Code's page](https://caveman.so/code#benchmarks).

### Every agent you use, in one window, on what you already pay for

Caveman Code is a Mac app that also runs in your browser. Claude Agent, Codex,
Cursor and Caveman, our own agent, work side by side in tabs and splits.

- **Bring what you already pay for.** Your Claude plan runs as Claude Agent and
  your ChatGPT plan as Codex, each on its own login. Caveman, our agent, takes
  any API key, coding plan or gateway you add. Free on your own keys and
  logins.
- **A Chief runs your chats.** Tell it what you want done. It starts the chats,
  answers their questions, reports back, and goes ahead on its own when you
  don't answer in time.
- **A guide to every pull request.** Written before you open it: a one-line
  TL;DR, the one call that is yours, then each change as before, after and
  why.
- **Agents that use a computer.** On a Cloud computer an agent drives a Linux
  desktop and a browser to check its own work. Watch it live, take over, hand
  back. On your Mac it works in your apps, in the background.
- **Close the lid. It keeps going.** Turn it on, and working chats continue on
  a Cloud computer when the lid closes. Add your other Mac with a one-time
  code and move a chat there too. Cloud computers are in beta.
- **Also in the app.** Automations that run by hand, on a schedule or on a
  GitHub event. A worktree for every chat. Agents that start agents and
  message each other. Plugins from Claude Code, Codex and Cursor. One memory
  that every agent reads. A terminal app is coming soon.

[All of Caveman Code](https://caveman.so/code). Caveman Code is not open
source.

## Caveman Code plans

A plan is credits: dollars, spent at each model's list price. On GLM-5.3 a
credit goes three times as far, and twice as far on DeepSeek Flash and GLM-5.3
Flash. Every other model counts at list price. Max $100 is the recommended
plan. Caveman Code is also free on your own keys and logins.
[Full pricing](https://caveman.so/pricing).

| Plan | A month | Credits every month | Up to, on GLM-5.3 |
| --- | --- | --- | --- |
| Pro | $20 | $20 | $60 |
| Pro $60 | $60 | $60 | $180 |
| Max $100 | $100 | $105 | $315 |
| Max $200 | $200 | $220 | $660 |

## Caveman Cloud

See and control every agent in one place. Soon, deploy the software factories
you build.

| | |
| --- | --- |
| For | The company paying for the agents |
| Runs | Hosted, or in your VPC on Enterprise |
| Works with | Coding agents, your own SDK calls, LiteLLM |
| Price | $0 for one person, no card needed. Team is $300 a month, with $300 of usage included, after a 7-day free trial. [Plans](https://caveman.so/pricing) · [Sign up](https://app.caveman.so/signup) |

### Deploy (coming soon)

Build your software factory in Caveman Code. Deploy it to Caveman Cloud. The
agents and automations you build on the Caveman Code harness become your
software factory. Soon you deploy it to Caveman Cloud, and it runs there
without your laptop.

1. **Build**, today: in Caveman Code, on the Caveman harness.
2. **Deploy**, coming soon: to Caveman Cloud, from the same account.
3. **Run**, coming soon: on Cloud computers, with its spend and traces beside
   every other agent's.

### Inside Caveman Cloud

A tour of the console, screen by screen. The screens show a demo workspace
with fictional data.

1. **Analytics. See what every agent costs.** One bill, split by agent, model
   and day: what your agents cost over the last 30 days, one colour per agent,
   and the money Caveman saved beside it. Green means verified.
2. **Traces. Every request, on the record.** Look up any call your agents
   made: which ones failed, and what each cost.
3. **Team. See who uses what.** Spend for each coding agent, across everyone
   on your team.
4. **Improvements. Caveman finds the waste.** Every fix it suggests says what
   it would save you, and is tried on your old requests first. On Autopilot,
   Caveman tries each change on a small share of traffic first, switches it
   back if quality slips, and asks you to keep it.
5. **Evals. Evals so easy, a caveman could do it.** Your own requests become
   tests, and every change gets graded before it ships. Pass a workload's
   quality gate and the higher Autopilot levels unlock.

[Try the demo](https://caveman.so/demo): a sample workspace, fictional data,
nothing to install.

### Govern

Set the limits. Keep the record. Soft and hard caps per project, rate limits
per key and workflow, and rules that mask or block what should never reach a
model. Owners and admins get an Inbox item and an email when a cap is crossed,
and every change lands in the audit log.

### What stays yours

- Your coding agents talk straight to your provider, with your key. Caveman
  gets a routing ask, never the request.
- On Enterprise, Caveman Cloud runs in your VPC, on your KMS and storage, with
  content sharing locked off.
- The local tools need no account and keep working if Cloud is unreachable.

## Already on Claude Code or Codex?

The open-source [Caveman CLI](https://caveman.so/caveman) makes the agent you
already use cheaper, in two commands. Free, Apache-2.0, runs on your machine.
It started as a joke that made agents talk like cavemen, then went to #1 on
Hacker News (904 points) and #1 on GitHub Trending, with more than 108,000
stars on GitHub. Now it runs inside 30+ coding agents.

```sh
npm i -g @caveman-ai/cli && caveman setup --install
caveman claude   # or codex · gemini · opencode
```

What it saves, and how each figure was measured, is on
[the Caveman CLI's page](https://caveman.so/caveman).

## Get started

```sh
npm i -g @caveman-ai/cli && caveman setup --install
```

That is the open-source CLI. For Caveman Code, [sign in](https://app.caveman.so/signup)
to use it in the browser, or [download the Mac app](https://caveman.so/download).
For your team's agents, [book a call](https://cal.com/caveman/chat) or
[try the demo](https://caveman.so/demo).

## For agents

- [Agent index and when-to-use guidance](https://caveman.so/llms.txt)
- [OpenAPI 3.1 specification](https://caveman.so/openapi.json)
- [API index](https://caveman.so/api)
- [OAuth 2.0 authorization server metadata](https://caveman.so/.well-known/oauth-authorization-server)

## Elsewhere

[Caveman Code](https://caveman.so/code) ·
[Models](https://caveman.so/models) ·
[Caveman Cloud](https://caveman.so/cloud) ·
[Caveman CLI](https://caveman.so/caveman) ·
[Pricing](https://caveman.so/pricing) ·
[Enterprise](https://caveman.so/enterprise) ·
[Compare](https://caveman.so/compare) ·
[Guides](https://caveman.so/guides) ·
[News](https://caveman.so/news) ·
[About](https://caveman.so/about) ·
[GitHub](https://github.com/JuliusBrussee/caveman)

## Before you start

### What is Caveman Code?

Caveman Code is an app for Mac and the browser that runs every coding agent you use in one window: Claude Agent, Codex, Cursor and Caveman, our own agent, with a Chief that runs your chats for you. It is free on the subscriptions and keys you already have. Sign in to use it in the browser on a Cloud computer, or download the Mac app.

[See Caveman Code](https://caveman.so/code)

### How can it cost less on the same model?

The harness around a model decides what every request carries and how many go out. Caveman, our own agent, runs on ours. On the same model, it finished Terminal-Bench 4 for 21% less than Codex, and day-long work across connected apps for 76% less than Claude Code. It did that on 31% fewer tokens than Codex per solved task, and 83% fewer than Claude Code per session. The runs and how we measured them are on the Caveman Code page.

[See how we measured](https://caveman.so/code#benchmarks)

### Can I use the subscriptions I already pay for?

Yes. Your Claude plan runs as Claude Agent and your ChatGPT plan as Codex, each through its own app and login; Caveman Code never copies a password or token. Caveman, our own agent, takes any API key, a coding plan such as Kimi Code or GLM Coding Plan, a gateway such as OpenRouter, or your own endpoint. Caveman Code is free on your own keys and logins.

[See what you can bring](https://caveman.so/code#app)

### What is Caveman Cloud?

Caveman Cloud is where your agents run and where you see and control them: spend, keys, policies and traces in one place. It is free for one person. Deploying the software factories you build with Caveman Code to Caveman Cloud is coming soon.

[See Caveman Cloud](https://caveman.so/cloud)

### Can I keep my coding agent?

Yes. The open-source Caveman CLI wraps supported agents including Claude Code, Codex, and Gemini CLI. Shorter replies also work in agents such as Cursor. Support for shorter replies and for context compression differs by agent, so check the setup guide for yours before installing.

[Check your agent's setup](https://caveman.so/guides/coding-agent-setup)

### Do I need a Caveman account?

Caveman Code and Caveman Cloud need an account, which you can open at app.caveman.so/signup. The open-source CLI, shorter replies and context compression work without one. You still need the credentials or subscription required by your model provider or coding agent. Adaptive model selection needs a free account.

[Set up the local tools](https://docs.caveman.so/caveman/quickstart)

### Will fewer tokens lower my bill?

It depends on how you pay. Fewer billable tokens can reduce metered API spend, but cached input, retries, and recovery calls affect the result. A flat monthly subscription may cost the same. Compare total cost per successful task and check provider billing before claiming savings.

[Measure cost per successful task](https://caveman.so/guides/measure-agent-cost)