One GraphQL API over all your data.
hugr classic puts databases, files and REST services behind one GraphQL schema, with roles and permissions, and an MCP server so AI assistants can query the data. It runs today, and it keeps running while it moves onto our new DuckDB core.
A data layer your applications and agents can share.
One schema
PostgreSQL, DuckDB, Parquet and other files, Iceberg and REST APIs appear as one GraphQL schema, joined across sources.
Aggregations built in
Group, aggregate and bucket data in the query itself, including spatial queries, without writing SQL.
Roles and permissions
Decide per role which types, fields and rows are visible, once, for every client.
MCP for AI assistants
An MCP server lets assistants explore the schema and query the data under the same permissions.
Nothing breaks. GraphQL moves to the new core.
hugr classic stays as it is
Available and stable. We fix problems; new features go into the new core.
GraphQL on the new DuckDB core
hugr’s GraphQL API moves onto our new DuckDB-based core, with the same ideas and a smooth path for existing schemas.
Running hugr today?
Tell us how you use it, and we will plan the move with you.