explaingit

helixdb/helix-db

★ 4,482RustAudience · developerComplexity · 3/5LicenseSetup · moderate

TLDR

An open-source database combining graph storage and vector similarity search in one system, built for AI agents and retrieval-augmented generation apps that need to search a knowledge base by meaning.

Mindmap

mindmap
  root((HelixDB))
    What it does
      Vector search
      Graph traversal
      AI data storage
    Query language
      HelixQL schemas
      Type-safe queries
      Compiled endpoints
    Clients
      TypeScript SDK
      Python SDK
      MCP protocol
    Use cases
      RAG pipelines
      AI agents
      Knowledge bases
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Build a RAG pipeline that searches documents by meaning and traverses entity relationships without separate vector and graph databases.

USE CASE 2

Store embeddings and connections for an AI agent so it can find contextually relevant information across your knowledge base.

USE CASE 3

Replace a separate vector store, graph DB, and document store with one HelixDB instance in your AI application stack.

USE CASE 4

Let AI agent frameworks automatically discover and query stored data using HelixDB's built-in MCP protocol support.

Tech stack

RustTypeScriptPythonLMDB

Getting it running

Difficulty · moderate Time to first run · 30min

Requires the Helix CLI to compile .hx schema files and push them to API endpoints before the database can be queried.

You can use and modify this software freely, but if you distribute a modified version you must also release it as open source under the same license.

In plain English

HelixDB is an open-source database written in Rust, designed to store and query data for AI applications. It combines two storage models in one system: a graph database, which stores data as nodes and connected edges to represent relationships, and a vector database, which stores numerical representations of content so that similar items can be found by meaning rather than exact text match. The intended use case is building AI agents and retrieval-augmented generation systems, where an application needs to search a knowledge base and pass relevant context to a language model. HelixDB is designed to replace what would otherwise require separate systems: a relational or document store for structured data, a vector store for similarity search, and a graph layer for traversing relationships between entities. Queries are written in HelixQL, a type-safe query language the project defines. Developers write schemas and queries in .hx files, then compile and push them to API endpoints using the Helix command-line tool. Client libraries for TypeScript and Python are provided. The database also includes a built-in text embedding function, so data does not need to be converted to vectors before insertion. HelixDB uses LMDB as its storage engine, which the README credits for low latency. Built-in support for MCP, a protocol used by some AI agent frameworks, allows agents to discover and traverse stored data without writing custom query code. The software is licensed under the AGPL. A managed hosted service is available for organizations that prefer not to run the database themselves.

Copy-paste prompts

Prompt 1
Show me how to define a HelixQL schema for a knowledge base that stores documents with vector embeddings and tag relationships.
Prompt 2
Write a TypeScript client for HelixDB that inserts text documents with auto-embedding and retrieves the top 5 most semantically similar results.
Prompt 3
How do I set up HelixDB locally and run my first vector similarity search using the Python client?
Prompt 4
I'm building a RAG pipeline, show me how to store text chunks in HelixDB, embed them automatically, and query for the most relevant ones.
Prompt 5
What are the AGPL license implications of using HelixDB in a commercial SaaS product that I don't open-source?
Open on GitHub → Explain another repo

← helixdb on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.