Knowledge graphs, vectors, full-text search, and transactional application data in one Rust engine — built for agents, GraphRAG, and company brains.
Vector search finds similar things. Graph traversal explains how they are connected. HelixDB combines both with BM25 full-text search and transactional application data, so your AI can retrieve richer context without stitching together multiple databases.
# Install the CLI
curl -sSL "https://install.helix-db.com" | bash
# Create and run a local project (requires Docker or Podman)
mkdir my-helix-app && cd my-helix-app
helix init local
helix start dev
# Run the generated example query
helix query dev --file examples/request.jsonLocal instances start in memory for a fast development loop. Use helix start dev --disk when you want data to persist across restarts.
Want an agent to bootstrap the project? Run helix chef to scaffold HelixDB, seed starter data, install the Helix skills and docs integration, and hand the build to Codex, Claude Code, or OpenCode.
Author strongly typed queries in your application and send them directly to a running instance — no separate query deployment step.
Helix Cloud runs the graph, vector, and text engines on durable object storage. A gateway routes requests to a dedicated writer and horizontally scaling readers, while tiered memory and SSD caches keep hot data close to compute.
- Object storage is the durable system of record
- Readers auto-scale independently with query load
- Serializable transactions provide full ACID guarantees
- High-availability deployments span multiple gateways and database nodes
Explore the architecture · Get started with Helix Cloud
Just use Helix.
