DevFlow Agent

• 

  Gemini AI diagnosing the root cause of pipeline failure with detailed fix steps

• 

  Gemini AI analyzing GitLab pipeline failures in real-time

• 

  DevFlow Agent homepage showing the 5-step CI/CD failure detection pipeline

• 

  Auto-created GitLab issue with AI diagnosis, labels and fix instructions

Inspiration

Every developer knows the pain of CI/CD pipeline failures. Getting woken up at 2am, staring at 500 lines of confusing error logs, spending hours debugging — this wastes valuable developer time every single day.

What it does

DevFlow Agent is an AI-powered DevOps assistant that automatically:

• Detects failed GitLab CI/CD pipelines
• Diagnoses root causes using Gemini AI
• Auto-creates GitLab issues with full analysis and fix steps
• Retries failed pipelines with one click

What used to take 2 hours now takes 30 seconds!

How I built it

• Gemini 2.5 Flash for AI-powered log analysis and diagnosis
• GitLab API + MCP for pipeline monitoring and issue creation
• Python + Flask for the backend server
• Google Cloud for AI infrastructure
• Render.com for deployment

Challenges I ran into

• Getting GitLab MCP integration working correctly
• Managing Gemini API quotas during development
• Deploying the Flask app correctly on cloud infrastructure

What I learned

• How to integrate GitLab MCP with AI agents
• Building multi-step AI agents that take real actions
• Deploying Python apps to cloud platforms

What's next

• Slack/Discord notifications for pipeline failures
• Auto-fix code using Gemini AI suggestions
• Support for GitHub Actions pipelines
• Pipeline failure analytics dashboard

Built With

• css
• flask
• gemini-ai
• gitlab-api
• gitlab-mcp
• google-cloud
• html
• javascript
• python
• render