PipelineIQ

• 

  Auto-generated Merge Request by PipelineIQ showing root cause analysis and suggested fix — zero manual debugging required

• 

  PipelineIQ AI agent detecting a failed pipeline, analyzing logs with Gemini AI, and automatically creating a fix MR in under 2 minutes

Inspiration

Every developer knows this pain. You push code, the pipeline fails, and you spend the next hour reading cryptic logs. PipelineIQ was born from this frustration.

What it does

PipelineIQ is an AI agent that watches your GitLab pipelines 24/7. The moment a pipeline fails, it analyzes the logs with Gemini AI, creates a fix branch automatically, and opens a Merge Request with a plain English explanation — all in under 2 minutes.

How we built it

• GitLab Duo Agent Platform for agent orchestration
• GitLab Webhooks to trigger on pipeline failures
• Gemini AI to analyze logs and generate fixes
• GitLab REST API to create branches and open MRs
• Node.js + Express for the webhook server
• Railway for cloud deployment
• Markdown-based memory system that learns from past fixes

Challenges

Getting Gemini to return consistent JSON required careful prompt engineering. Building memory without a database — solved with simple markdown files.

What we learned

AI agents are most powerful when they react to real triggers. Memory doesn't need a database — simple files work perfectly.

What's next

• Fetch actual job logs for deeper analysis
• Support multiple repos
• Add Slack notifications
• Build a dashboard

Built With

• express.js
• gemini-ai
• gitlab-duo-agent-platform
• gitlab-rest-api
• javascript
• node.js
• railway