Code never leaves your machine
Reviews run against your local Ollama instance—ideal for air-gapped or highly sensitive repositories.
Feature · Local AI
Fully offline reviews with Ollama
Connect ThinkReview to Ollama on your machine—air-gapped and privacy-sensitive teams get AI code review without sending patches to the cloud.
Chrome · Firefox · Edge · Brave
Why teams use this
Built to fit your existing git workflow—no new platform to adopt.
Your choice of local models
Use Llama, Qwen Coder, DeepSeek, and other Ollama-supported models—swap as your hardware allows.
Cloud or local, same UI
Pick Ollama next to Cloud AI in extension settings—same PR sidebar workflow on GitHub, GitLab, Azure DevOps, and Bitbucket.
Open-source friendly
ThinkReview is open source—inspect the extension and run local inference end to end.
How it works
Get from install to first review in a few minutes.
1
Install and allow CORS
Run Ollama locally and set OLLAMA_ORIGINS so the browser extension can reach it.
2
Select Ollama in settings
In the ThinkReview extension, choose Ollama as the provider and pick a pulled model.
3
Review a PR offline
Open a pull or merge request—analysis runs on your local models.
See it in context
Related
Explore more features, platforms, and documentation.
Review PRs with local models
Install ThinkReview, connect Ollama, and keep code on your machine.