AI-DEVOPS-AGENT
Link to open source: https://github.com/BhavyaBhardwaj807/AI-DevOps-Agent.git
The system is designed as an AI-driven DevOps orchestration layer rather than a standalone chatbot. The React/Vite dashboard acts as the user-facing control layer, where developers provide a repository and receive real-time engineering insights. A FastAPI backend handles the application logic and keeps all external integrations away from the frontend.
At the core is Swytchcode, which acts as the agent's execution and integration layer. Instead of directly connecting the application to individual DevOps APIs, the backend invokes Swytchcode tools to interact with services such as GitHub, Jira, and Netlify. This makes the architecture modular, so additional DevOps tools can be plugged into the agent without restructuring the application.
For repository intelligence, the agent retrieves live GitHub activity through Swytchcode and passes the resulting data to GPT-4o-mini via AICredits. The AI converts raw repository activity into structured intelligence — including activity levels, risk assessment, key changes, potential issues, and actionable recommendations — which is then presented through the dashboard.
The architecture is designed around a Detect → Analyze → Act workflow: the agent observes engineering activity, uses AI to understand potential risks and improvements, and provides a path toward executing those recommendations through DevOps tools such as Jira and Netlify.
Architecture flow:
React/Vite → FastAPI → Swytchcode → GitHub → GPT-4o-mini → DevOps Insights → Jira/Netlify Actions
