fastapi mcp langgraph template

Report Abuse

Basic Information

This repository is a developer-focused template for building and deploying agentic orchestration systems. It bundles a FastAPI backend, MCP client/server integration and LangGraph orchestration to help teams prototype multi-component agent workflows with streaming and persisted chat state. The template includes database integration patterns using SQLModel with PostgreSQL and PGVector, RBAC via Supabase, and observability hooks with Langfuse. It provides Docker Compose configurations for development and production, an optional VSCode devcontainer workflow for debugging, and architecture documentation including an Inspector component that communicates with MCP servers over SSE. The README documents setup, build scripts for example MCP servers, environment variable management, and planned production features like metrics, auth, and CI/CD. The project is intended as a starting point for engineers building scalable, customizable agent platforms.

Links

App Details

Features
MCP and LangGraph integration for standardized context delivery and customizable agentic orchestration including native streaming and persisted chat/state management. FastAPI-based Python backend using SQLModel and Pydantic for ORM and validation. Relational Postgres storage with PGVector for vector embeddings and Supabase for DB RBAC. Langfuse instrumentation for LLM observability and metrics. Docker Compose and Dockerfile builds for development and production, plus a community build script for example MCP images. Nginx reverse proxy example and VSCode devcontainer support for debugging. Documentation and modular docs for MCP, LangGraph, Supabase, Langfuse, and Grafana stack. Planned CI/CD and infrastructure IaC notes mentioning Terraform, Prometheus, Grafana, Auth0, and container registry deployment.
Use Cases
The template accelerates development of multi-agent and agentic orchestration projects by providing a concrete, opinionated stack and patterns for common needs. Developers get an integrated FastAPI API server, database models, vector storage, and observability hooks out of the box so they can focus on agent logic rather than infrastructure plumbing. Docker Compose files and build scripts make local development and reproducible production images straightforward. The Inspector and architecture diagrams clarify service interactions and SSE-based communication to MCP servers. Built-in examples and docs reduce onboarding time, and the template explicitly avoids LLM provider lock-in while outlining production additions like monitoring, auth, and CI/CD for teams preparing to scale.

Please fill the required fields*