Inspiration
Developers spend too much time understanding complex systems instead of building them. We wanted to make architectural knowledge instantly accessible through AI.
What it does
ArchLens is an AI Software Architect in Slack that answers questions about your codebase, dependencies, APIs, pull requests, and system architecture.
How we built it
We built ArchLens around an automated repository indexing pipeline.
GitHub webhooks trigger indexing whenever code changes. The indexer parses source code, OpenAPI specifications, documentation, imports, package dependencies, service calls, and Git history.
These relationships are stored in a Neo4j knowledge graph alongside PostgreSQL for structured metadata and Redis for caching.
An MCP server exposes this architectural context to the OpenAI Responses API, enabling the model to retrieve accurate, architecture-aware information before generating responses.
Developers interact with the system entirely through Slack, creating a seamless workflow without switching tools.
Our frontend is built with Next.js, TypeScript, Tailwind CSS, and React Flow for future architecture visualization, while the backend uses Node.js and Express.
Log in or sign up for Devpost to join the conversation.