Skip to main content
Currently in Closed Beta — Code Review is available for GitHub repositories only. There is no additional charge during the beta period. Pricing details will be announced before general availability.

Overview

Unblocked Code Review uses AI to review your pull requests and catch real issues before they ship. It focuses on correctness and learns your team’s conventions over time. As you connect more context to your Unblocked workspace, reviews become increasingly precise and helpful. The AI reviewer analyzes code changes for potential bugs, logic errors, security vulnerabilities, and adherence to best practices. It supports all major programming languages and integrates seamlessly with your existing GitHub workflow.

Setting Up Code Review

To enable Code Review in your Unblocked team, navigate to Settings > Code Review. Then select the repositories you want to enable. Select Repositories

Custom Path Exclusion Patterns

You can configure custom path exclusion patterns to exclude specific files or directories from review. These patterns use glob syntax and are applied when processing pull requests. Common exclusion patterns include *.md to exclude all Markdown files at any depth, src/test/** to exclude all files under a specific directory from the repository root, and package-lock.json to exclude specific files. You can also exclude entire directories like node_modules/** or file types like *.generated.js. Exclusion Patterns

How Code Review Works

Once enabled, Unblocked automatically reviews each open pull request when it’s created or updated. The AI analyzes code changes for potential bugs and issues, then posts inline code comments directly in your PR when issues are found. Reviews typically complete within three minutes, depending on the size of the changes. Binary files are automatically skipped to focus on reviewable code.

Generated and Vendored Files

Unblocked automatically skips generated and vendored files so reviews stay focused on the code you actually maintain. It reads your repository’s .gitattributes and honors the same GitHub Linguist markers you may already use. If a file is marked as generated or vendored, Unblocked leaves it out, reducing noise and avoiding comments on machine-produced or third-party code. This includes common patterns like compiled assets, dependency files, and auto-generated documentation. If you already use Linguist to tidy language stats or collapse large diffs on GitHub, you don’t need to change anything—Unblocked follows those signals. Otherwise, you can add simple rules like:
generated/** linguist-generated
vendor/**    linguist-vendored
Learn more about Git attributes and GitHub Linguist overrides.

Draft Pull Requests

By default, Unblocked does not review draft PRs. This allows you to work without receiving automated feedback until you’re ready. If you prefer to receive code reviews on draft PRs, you can enable this in your user preferences by turning on the Review my draft pull requests setting.

User Preferences

Each developer can customize their code review experience from their personal preferences. To update these settings, go to Settings > Personal Settings > Preferences.
  • Receive automated code reviews as comments on my pull requests — You can receive automated code reviews as comments on your pull requests. Turn this on or off to control whether you get automated feedback. When disabled, you won’t receive automated comments, but team-level reviews still run for other team members.
  • Review my draft pull requests — If you want early feedback during development, enable review of your draft pull requests. Unblocked will review drafts in addition to regular PRs.
  • Only comment when issues are found — To reduce notifications, you can choose to only receive comments when issues are found. In that mode, Unblocked doesn’t post summary comments when everything looks good; it only comments when it identifies issues.

Providing Feedback

Provide feedback on code review comments by reacting with 👍 or 👎 emojis or by replying directly with additional context. Unblocked uses this feedback to refine its understanding of your team’s preferences and improve the quality of future reviews. Your feedback helps tune Unblocked to better understand your team’s coding standards and preferences. Positive feedback reinforces good suggestions, while negative feedback helps the system learn what types of comments aren’t helpful for your codebase. To get the most out of feedback, add a brief explanation when a comment isn’t helpful. That context helps the AI understand nuanced preferences and reduces similar comments in the future.