ProPR and Qodo: the work before and after the review

Qodo, which replaced Qodo Merge, reviews pull requests on GitHub, GitLab, Bitbucket, Azure DevOps, and Gerrit, and commits fixes. ProPR reviews too, and runs the rest of the loop on your server, plan to follow-up commit.

ProPR
vs
Qodo

When the loop starts

Qodo starts when a pull request exists. ProPR creates the pull request from an approved plan and keeps changing it from GitHub feedback.

Qodo pathExisting pull request → /agentic_review → remediation PR or pushQodo starts when the pull request exists.
ProPR pathPlan → isolated workspace → GitHub PR → follow-up commitProPR opens the pull request from an approved plan and keeps changing it.

Qodo Merge has become Qodo, a proprietary AI code review platform driven by PR commands such as /agentic_review, /ask, and /fix, with @qodo chat on the pull request. Its remediation agent applies a finding on a child branch and opens a remediation PR, or pushes straight to the source branch when configured to. Pricing is usage-based from a $30 per month Pro Team base plus credits, with Enterprise adding on-premises deployment. The open-source engine behind it, PR-Agent, left Qodo's hands in April 2026: it now lives in the community-run The-PR-Agent organization, under the MIT license, with a governance committee and a planned donation to a foundation.

ProPR is Apache 2.0 and self-hosted end to end. It turns an approved plan into an isolated agent run, opens the PR, and keeps it moving as review comments become follow-up commits, with logs and usage on your infrastructure. Run Qodo on the PRs ProPR opens if you want its review alongside /review.

A review scored 6 with two issues, then the /fix run that applies finding F1 and suggestion S2 and reruns the checkout tests
A review scored 6 with two issues, then the /fix run that applies finding F1 and suggestion S2 and reruns the checkout tests.

Review automation versus implementation follow-through.

QuestionQodoProPR
LicenseProprietary; PR-Agent engine is MIT, community-maintainedApache 2.0
DeploymentCloud; on-premises on Enterprise; PR-Agent self-hostsSelf-hosted, every installation
Review and automate existing PRsYesYes
Plan and implement before the PR existsNoYes
Turn review feedback into commitsYes: committable suggestions and a remediation PR or pushAny plain comment or /fix
Git hostsGitHub, GitLab, Bitbucket, Azure DevOps, GerritGitHub only

Qodo fits best when...

  • You need review, PR chat, and remediation on PRs your developers already write.
  • Your repositories span GitLab, Bitbucket, Azure DevOps, or Gerrit as well as GitHub.
  • Implementation stays with the team or another tool.

ProPR fits best when...

  • One system takes the change from plan to PR on your own server.
  • Review feedback becomes follow-up commits: plain comments for human requests, or /review then /fix for AI review findings.
  • Agent choice, logs, and cost records live in your own deployment.

Verified November 2026: Qodo code review docs, Qodo remediation, Qodo git providers, Qodo pricing, PR-Agent handover, PR-Agent repository.

Use Qodo with ProPR.

Stays in Qodo

  • /agentic_review, /ask, and @qodo chat on every pull request.
  • Remediation pull requests or pushes for Qodo's own findings.
  • Review across GitLab, Bitbucket, Azure DevOps, and Gerrit, where ProPR does not reach.

Moves to ProPR

  • The plan and the implementation before the pull request exists, on your server with the agent you choose.
  • Follow-up commits for human review comments, in plain language.
  • Task history, logs, and usage for every run.

Wire them together

Install ProPR and keep Qodo installed

Run propr setup (quickstart) or pick a ProPR Hosted plan, add the repository, and sign in your agent. Qodo keeps its GitHub installation.

Let ProPR open the pull request

Add the AI label to an issue and approve the plan. Qodo reviews the pull request ProPR opens, as it does any other.

Answer in plain language

Write what you want changed as a plain comment and ProPR pushes the commit. Keep Qodo's remediation for its own findings, or run /review and /fix for a second model's. The follow-up playbook shows one comment becoming a commit.