AI-Powered GitLab Merge Request Code Reviewer

Unlock Automated Code Quality with this AI Agent

This AI Agent acts as an automated code reviewer for your GitLab merge requests. When triggered by a specific comment (e.g., +0) on an MR, it fetches the code changes, sends them to an OpenAI LLM for analysis, and posts a structured review back into the MR's discussion thread. This helps streamline your development workflow, catch potential issues early, and maintain high code quality standards with less manual effort.

Key Features & Benefits

• Automated GitLab MR Review: Triggers on specific merge request comments in GitLab (customizable, e.g., +0).
• AI-Powered Code Analysis: Leverages OpenAI's advanced language models to scrutinize code diffs for potential issues, style inconsistencies, or areas for improvement.
• Structured Feedback: Delivers reviews directly in the MR discussion, including a decision (accept/reject), a score (0-100), and detailed comments in Markdown format, as per your configured prompt.
• Customizable AI Persona: Tailor the AI reviewer's behavior and feedback style by modifying the system prompt in the 'Basic LLM Chain' node.
• Intelligent Diff Processing: Parses code changes and prepares them for AI analysis, skipping irrelevant file changes like renames or deletions.
• Saves Developer Time: Reduces the burden of initial manual code reviews, allowing developers to focus on more complex tasks and architectural decisions.
• Improves Code Consistency: Helps enforce coding standards and best practices automatically across projects.
• Accelerates Development Cycles: Provides faster feedback on code changes, leading to quicker iterations and deployment.

• An n8n instance (Cloud or self-hosted).
• GitLab account and a Personal Access Token (PAT) with api scope to read merge request data and post comments/discussions.
• OpenAI API Key with access to a suitable model (e.g., gpt-3.5-turbo, gpt-4).

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. Webhook Configuration:
   • Open the 'Webhook' node. Copy its 'TEST URL' (or 'PRODUCTION URL' once activated).
   • In your GitLab project, navigate to Settings > Webhooks.
   • Add a new webhook: Paste the copied n8n Webhook URL into the 'URL' field. Under 'Trigger', select 'Comments' and ensure 'Merge request events' is checked. Add the webhook.
4. GitLab API Integration:
   • In the 'Get Changes1' (HTTP Request) node:
     • Update the GitLab instance URL in the 'URL' parameter if you are using a self-hosted GitLab (e.g., https://your.gitlab.instance/api/v4/...).
     • Under 'Authentication', select 'Header Auth'. Add a new credential or select existing credentials for GitLab. The node expects a header named PRIVATE-TOKEN with your GitLab PAT as its value.
   • In the 'Post Discussions1' (HTTP Request) node:
     • Similarly, update the GitLab instance URL if necessary.
     • Configure 'Header Auth' with the same GitLab PAT credentials.
5. OpenAI Configuration: In the 'OpenAI Chat Model1' node, select or create your OpenAI credentials by entering your OpenAI API Key.
6. Customize AI Reviewer (Highly Recommended):
   • In the 'Basic LLM Chain1' node, review and edit the 'System Message' in the 'Messages' parameter. This prompt dictates the AI's persona, review style, scoring, and output format. Tailor it to your team's specific needs and coding standards (the default prompt provides a good starting point for a 'senior programming expert Bot').
7. Customize Trigger Condition (Optional):
   • In the 'Need Review1' (IF) node, inspect the conditions. By default, it triggers if a comment on an MR is exactly +0. You can change this value or add more complex logic to suit your review workflow.
8. Save any changes to the nodes.
9. Activate the workflow in your n8n instance.
10. Test by creating a merge request in your configured GitLab project and posting the trigger comment (e.g., +0) on it.