HelixDB

HelixDB is an open-source graph-vector database for AI memory that combines property graph structures with native vector and full-text search in one Rust engine.

HelixDB is an open-source graph-vector database designed for AI agent memory and RAG applications. It functions by allowing developers to store data as a labeled property graph where vector, text, and secondary indexes operate as access paths over nodes and edges. This architecture eliminates the need for keeping separate graph, vector, and search stores in sync. The system runs on object storage and utilizes a Rust engine to perform prefiltered search operations, where graph traversals define the candidate set before ranking occurs within a single ACID transaction. Official SDKs are available for Rust, TypeScript, Python, and Go to handle data and queries, which are sent via JSON. It supports deployment via embedded processes, Docker containers, or as a managed cloud service.

HelixDB at a glance

Best for: Software engineer
Runs on: macOS, Linux, Windows, Docker, Podman
Pricing: Free for self-hosting and embedding. Helix Cloud starts at $5/month for 200K reads and 100K writes; Growth plan offers dedicated clusters billed by hour.
Released: November 2024
Category: AI coding tool
Website: helix-db.com

Key features

Native vector and BM25 full-text search
Graph-based data traversal and querying
ACID transactions with serializable snapshot isolation
Support for in-memory, disk, and object storage backends
Typed SDKs for Rust, TypeScript, Python, and Go
Prefiltered search capabilities within graph traversals
Single-engine architecture for unified data management

Use it for

Building long-term memory for AI agents
Grounding AI application answers in specific user records
Connecting code, documentation, and project data in a graph
Retrieving context for domain-specific AI queries

Alternatives to HelixDB

seekdb — Unifies relational, vector, and text data for production-grade AI agents.
Helix Foundry — Connects company data into a local ontology with local-first AI and no cloud accounts.
Helixharness — It brings your team engineering DNA and verified project lessons directly into AI development.
Open Glean — Provides agents with structured state and relationship-based context rather than just similarity.
CodeGraph — It eliminates file crawling, reducing AI agent tool calls by 88% and cutting costs.
Codebase Memory MCP — High-speed indexing and persistent knowledge graphs for AI coding agents.
Qdrant — Fast filtered vector search, self-hosted or in a managed cloud
XERJ — It answers agent questions from an index instead of wasting tokens on context windows.
zvec — It is a pure local, serverless vector database providing fast hybrid search.
Infinity — Achieves 0.1 milliseconds query latency on million-scale vector datasets.
Glider — It combines crash-safe S3 durability with fast local SSD caching for vector search.
Weaviate — Stores data objects and vectors for semantic and hybrid search capabilities.

All HelixDB alternatives