Getting started with Pullora

Pullora reviews your Pull Requests with AI — posting inline comments on the exact changed lines plus a concise summary. Setup takes about five minutes.

1. Sign in

Sign in with Google, Microsoft, GitHub, GitLab or Bitbucket — whichever your team already uses. There are no passwords to manage. Your workspace is created automatically, and you can invite teammates later on a Team plan.

2. Connect your git provider

Pullora runs one review engine behind four providers. Connect whichever you use from Dashboard → Repositories; you can connect more than one to the same workspace.

GitHub

Click Connect GitHub and install the Pullora GitHub App on the repositories you want reviewed. Pullora uses least-privilege, short-lived installation tokens — never personal access tokens — and you can change the repository selection any time from GitHub settings.

GitLab

Click Connect GitLab and authorise the OAuth app. Nothing else is needed — your accessible projects are synced automatically, and merge requests are reviewed the same way pull requests are.

Bitbucket Cloud

Click Connect Bitbucket and enter your workspace name before authorising — Bitbucket removed workspace discovery from its API, so we have to ask. It is the segment in any repository URL: bitbucket.org/acme-team/my-repo means the workspace is acme-team.

Azure DevOps

Azure DevOps connects with your organisation URL and a Personal Access Token. Create a PAT with Code (Read & write) and Service hooks (Read & write) scopes, then paste it with your organisation URL (for example https://dev.azure.com/your-org). The token is encrypted at rest and is the one case where Pullora stores a long-lived credential, because Azure DevOps offers no app installation model.

3. Run your first review

There are two ways:

• Paste a PR URL — go to Dashboard → Review PR, paste something like https://github.com/your-org/your-repo/pull/123, pick a mode, and submit. Tip: check Dry run the first time — findings appear only in your dashboard, and nothing is posted to the pull request until you choose to publish.
• Enable automatic reviews — in Dashboard → Repositories, toggle Auto Review on. Every new or updated PR is reviewed automatically; new commits get an incremental review of only the new changes.

4. Read the results

Each finding shows a severity (Critical / Major / Minor / Suggestion), the file and line, an explanation of what can go wrong, a suggested fix, and a confidence score. Only high-confidence findings become inline comments; the rest stay in the summary. Use the Helpful / Not helpful / False positive buttons — findings you dismiss are never repeated on later runs.

5. Tune it to your team (optional)

• Commit a .ai-review.yml to your repository to set review depth, ignored paths, and your own plain-English rules.
• Or adjust per-repository settings in the dashboard: minimum severity/confidence, category toggles, branch filters, and custom instructions.
• Choose a default review mode: Quick (from 0.5 units), Standard (1), or Deep (2).

Questions? Contact us — or see the pricing page for plan limits.