PR-Agent
AI code review agent that describes, reviews and suggests improvements on pull requests via comments, CI or the CLI.
Run /describe, /review, /improve and /ask on GitHub, GitLab, Bitbucket, Azure DevOps or Gitea.
PR-Agent is an open-source AI code review agent that describes, reviews and improves pull requests. It runs from a PR comment, a webhook, your CI or the terminal. It is now a community-maintained legacy project of Qodo, separate from Qodo's commercial code review product.
What it is
You trigger commands such as /describe, /review, /improve and /ask on a pull request and the agent posts results as comments. The docs show an example reviewer guide covering review effort, tests and security concerns. It works with GitHub, GitLab, Bitbucket, Azure DevOps and Gitea, and you supply the model API key.
Who it's for
It suits engineering teams that want AI review inside their own CI and are happy to bring their own OpenAI or other model key. It does not suit teams wanting a vendor-supported product: the README says it is community-maintained and distinct from Qodo's primary offering.
Key features
Slash commands on PRs
/describe, /review, /improve and /ask work from a pull request comment.
GitHub Action setup
A workflow file adds automated review on opened and updated pull requests.
Local CLI mode
pip install pr-agent and run it against a PR URL from your terminal.
Multiple git hosts
Docs cover GitHub, GitLab, Bitbucket, Azure DevOps and Gitea.
Docker image
The pragent/pr-agent image runs the same review without a Python install.
Reviewer guide output
Reviews flag effort, tests and focus areas such as unbounded backoff.
Up and running in 3 steps
Our honest verdict
What's great
- Runs inside your own GitHub Actions or webhook, so diffs go only to the model provider you configure.
- Supports five git hosting platforms, not only GitHub.
- Install options include a pip package with optional extras for each git provider.
- Commands like /ask let reviewers question the agent about a pull request.
Where it falls short
- You must supply and pay for your own LLM API key, such as an OpenAI key, so cost scales with the size and number of pull requests.
- The README calls it a community-maintained legacy project of Qodo, so do not expect vendor support or the features of Qodo's paid review product.
- Webhook installs for GitLab, Bitbucket or Azure DevOps need you to set up a server and credentials following the docs.
- Release notes warn of breaking changes, such as settings that repository files can no longer override, so upgrade carefully.
Pricing & license
PR-Agent is MIT-licensed and free. The docs at https://docs.pr-agent.ai show no paid plan for the project, so the cost is the LLM API usage you pay your model provider plus any CI minutes.
Open remote AI-training roles
767 open remote AI-training roles, updated November 14, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Is PR-Agent the same as Qodo?
No. The README says it is a community-maintained legacy project of Qodo and distinct from Qodo's primary AI code review offering.
Which model does it use?
The quick start uses an OpenAI key. The docs cover other model and git provider setup.
Where does the repo live now?
The release and docs pages show it under the The-PR-Agent organization, and the qodo-ai/pr-agent URL redirects there.
