The article introduces HelixDB, a Rust‑based database that combines graph, vector and key‑value capabilities for AI workloads, describing its CLI, SDKs and cloud offering.
HelixDB promises a single engine that merges graph, vector and key value stores for AI memory and knowledge graph use cases. The project is written in Rust and aims to replace a stack of separate databases with one service.
The announcement describes a CLI that can spin up a local instance, a chef command that scaffolds a project and seeds data, and SDKs for Rust and TypeScript that let developers send queries as dynamic POST requests. What stands out is the tight integration of graph traversal, vector search and simple key value operations behind a single query interface. The system also supports relational style tables, documents and a KV store, all accessible through the same API.
The default storage mode runs only in memory, so data disappears when the process stops unless a disk flag is used. The documentation notes that the in‑memory mode is intended for quick prototyping, not for production persistence. Installation is a one line curl command that writes the CLI to the path ~/.local/bin/helix. After installation the quick start uses helix chef to create a project, start a local node on port 6969 and run a sample query file. A manual path lets developers initialise a helix.toml file, launch a background container and send queries with the provided SDKs.
website docs discord X/twitter
HelixDB Cloud adds an object storage backend, full ACID transactions, a single writer with auto scaling readers and high availability across three gateways. The cloud service is managed through a separate auth flow that pulls a cluster ID and gateway URL into the local configuration. 
Queries are authored with a Rust or TypeScript DSL and sent straight to a running instance as dynamic requests against POST /v1/query. The Rust SDK lets you register functions with the #[register] attribute and call them through a client that defaults to http://localhost:6969. The TypeScript SDK follows a similar pattern with defineParams and dynamicJson helpers. Both SDKs produce the same JSON AST and can be used without a build step.
The post ends with a note that the project is open source on GitHub and that commercial support is available through the managed cloud offering.
Comments
Please log in or register to join the discussion