HelixDB
Graph-vector database for AI
HelixDB is an open-source graph-vector database that brings structure to unstructured data for RAG and AI applications, combining vector meaning with graph relationships.
- Founders
- George Curtis, Xavier Cochran
- YC batch
- P25
Company posts
Summaries from the company’s own website.
- How We Used PlanetScale MCP to Cut Query Latency ↗
HelixDB says it used PlanetScale Insights and the PlanetScale MCP server to identify an expensive cursor scan in its control-plane queries, trace it to two backend workers, and test fixes against live schema and metrics. The company says a smaller predicate rewrite let MySQL and Vitess use the existing composite index, reducing wasted database work and lowering query latency percentiles.
- Introducing HelixDB Query Insights ↗
HelixDB says it launched Query Insights in HelixDB Cloud to help users understand why queries are slow by combining latency graphs, planner findings, index information, and recommendations in one view. The post also says these insights are available in the dashboard and through its MCP server for agent-based troubleshooting.