Repolode
Submit
docker

docker/genai-stack

A Docker-based starter stack for building GenAI apps with LangChain, Neo4j, and Ollama, including support bot, PDF Q&A, and API demos.

View on GitHub
Stars
5,423
Forks
1,236
Open issues
99
Last commit
12 days ago

What is it?

GenAI Stack is a starter kit for building generative AI applications with Docker Compose. It wires together LangChain for orchestration, Neo4j as a database that stores both a knowledge graph and vector embeddings, and Ollama for running local LLMs, so developers can start from a working multi-service setup instead of assembling the pieces themselves.

The repository ships five demo applications: a support agent bot that compares answers with and without retrieval-augmented generation (RAG), a loader that imports Stack Overflow data into the database, a question-and-answer app for local PDFs, a standalone HTTP API with streaming and non-streaming endpoints, and a Svelte front-end that calls that API. Configuration is handled through a `.env` file that selects the LLM and embedding model.

LLMs can be served by a local Ollama instance or by OpenAI, AWS Bedrock (Claude v2), or Google models, with separate profiles for CPU and GPU Ollama containers on Linux.

Key features

  • 01Docker Compose setup that runs the LangChain, Neo4j, and Ollama services together
  • 02Neo4j stores a knowledge graph and vector embeddings for retrieval
  • 03Switch between local Ollama models and OpenAI, AWS Claude, or Google embedding models via environment variables
  • 04Support bot that compares RAG-enabled and pure LLM answers
  • 05Standalone HTTP API with SSE streaming endpoint
  • 06Svelte front-end built with Vite and Tailwind

Use cases

  • Prototyping a retrieval-augmented chatbot over your own documents or data
  • Learning how to combine a knowledge graph with vector search in an LLM app
  • Chatting with a local PDF using embeddings and a local model
  • Serving an LLM-backed question-answering API to a separate front-end

Quick start

  1. 1

    Create a .env file from the template env.example

  2. 2

    On macOS or Windows, install and start Ollama (ollama serve) in a separate terminal

    ollama serve
  3. 3

    Start the full stack

    docker compose up
  4. 4

    Open the Support Bot at http://localhost:8501

Things to know

  • Requires Docker and Docker Compose, and Docker Desktop 4.24.x has a known performance issue that needs an upgrade.
  • Local Ollama models on macOS and Windows need Ollama installed separately; GPT-4 or Claude options need paid API keys.
  • Some setups require a GPU for the Linux GPU profile.

In collections

Private AI knowledge bases and RAGRun LLMs locally

Topics

#genai#langchain#neo4j#ollama#docker#rag#vector-database#knowledge-graph

FAQ

What is genai-stack?
A Docker-based starter stack for building GenAI apps with LangChain, Neo4j, and Ollama, including support bot, PDF Q&A, and API demos. GenAI Stack is a starter kit for building generative AI applications with Docker Compose. It wires together LangChain for orchestration, Neo4j as a database that stores both a knowledge graph and vector embeddings, and Ollama for running local LLMs, so developers can start from a working multi-service setup instead of assembling the pieces themselves.
Is genai-stack free to use? What license does it have?
genai-stack is open source under the CC0-1.0 license. Check the license terms before using it in commercial products.
Is genai-stack still maintained?
Actively maintained. The last commit to genai-stack was on Oct 26, 2026.
What can I use genai-stack for?
Prototyping a retrieval-augmented chatbot over your own documents or data Learning how to combine a knowledge graph with vector search in an LLM app Chatting with a local PDF using embeddings and a local model Serving an LLM-backed question-answering API to a separate front-end
How do I install genai-stack?
Create a .env file from the template env.example On macOS or Windows, install and start Ollama (ollama serve) in a separate terminal: ollama serve Start the full stack: docker compose up Open the Support Bot at http://localhost:8501
Where can I get genai-stack?
The source code and installation instructions are on GitHub: https://github.com/docker/genai-stack

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects