Question 1

How can it be free?

Accepted Answer

Freebuff is supported by text ads.

Question 2

What models do you use?

Accepted Answer

Freebuff runs on the best open-source models available. In full mode, you can choose from:

- DeepSeek V4 Pro: smartest. Its API collects data for training.
- MiMo 2.5 Pro: smartest and multimodal, but slower.
- Kimi K2.6: balanced and multimodal.
- DeepSeek V4 Flash: most efficient. Its API also collects data for training.
- MiMo 2.5: multimodal.
- MiniMax M3: smartest unlimited model, multimodal. Its API collects data for training.

Limited mode uses DeepSeek V4 Flash and MiMo 2.5.

Also, Gemini 3.1 Flash Lite handles file finding and research.

Question 3

Which countries is Freebuff available in?

Accepted Answer

All countries. Freebuff is available in "full" or "limited" mode. The following countries have full access:

United States, Canada, United Kingdom, Australia, New Zealand, Norway, Sweden, Netherlands, Denmark, Germany, France, Italy, Spain, Portugal, Finland, Belgium, Luxembourg, Liechtenstein, Switzerland, Austria, Singapore, Malta, Israel, Ireland, and Iceland.

If you are outside those countries or using a VPN, Freebuff still works in limited mode.

Question 4

What is limited mode?

Accepted Answer

Limited mode lets you use Freebuff outside the full-access countries, or while using a VPN. It includes DeepSeek V4 Flash and MiMo 2.5, with 5 one-hour sessions per day.

Question 5

Are you training on my data?

Accepted Answer

No. We do not share your data with third parties that would train on it or use it for another purpose, unless you choose a model clearly labeled as 'Collects data for training'.

Question 6

What data do you store?

Accepted Answer

We don't store your codebase. We only collect minimal logs for debugging purposes.

Question 7

What else is cool in Freebuff?

Accepted Answer

Freebuff comes with 9 specialized subagents:
- file-picker finds relevant files across your codebase
- code-reviewer gives critical feedback on your changes
- browser-use lets the AI control a real browser to test your app
- thinker does deep reasoning
- and more.

After every response, it generates 3 clickable follow-up suggestions so you always know what to do next.

For big tasks, try the commands /interview → /plan → (implement) → /review to go from idea to polished code.

The free coding agent for your terminal

Up and running in under a minute

Questions, answered

freebuff