System Map

How SharpSana is organized

Use this page when you need the technical shape of the product quickly: major surfaces, domain objects, service responsibilities, and the mental model an engineering agent should use.

Product Shape

Next.js Web App

Dashboard, project workspace, settings, API keys, docs, and onboarding.

Supabase Data Model

Projects are the center. Tasks, columns, sprints, insights, materials, threads, and roadmap items attach to project context.

Cloud Run Services

Analysis, chat, materials indexing, interview processing, roadmap generation, sprint generation, repo indexing, repo search, GitHub integration, and MCP.

CLI + MCP Clients

The CLI binds a local repo to a SharpSana project and configures IDE MCP clients to use that context.

Core Architecture

Web App / CLI / IDE MCP Client
        |
        v
   SharpSana Services
        |
        v
     Supabase
        |
        v
GitHub / Storage / Gemini / Groq / Polar

Core Entities

ProjectMaterialInterviewInterviewBatchInsightTaskColumnSprintRoadmapItemThreadRepositoryIndexAPI Key

Service Responsibilities

cloud-run-project-chat: main project assistant and orchestration layer
cloud-run-project-analysis: initial AI analysis of project memory
cloud-run-materials-indexer: file and link extraction, chunking, embeddings
cloud-run-interview-processor: transcription, analysis, interview chat
cloud-run-repository-indexer: repo structure summary in Supabase
cloud-run-repo-search: code and symbol search
cloud-run-sprint-generator and cloud-run-roadmap: planning outputs
cloud-run-mcp-context: IDE-facing MCP tool and resource server

Agent Mental Model

SharpSana has three linked jobs: collect context, reason over context, and turn context into execution artifacts and IDE-usable guidance.

Next: Getting Started