Vibe Coding Discoverpublisher
View on GitHubPublisher is the open-source analytics engine for Malloy. It lets you define data models once — and use them everywhere.
Malloy Publisher is an open-source analytics engine that serves Malloy semantic data models over MCP and REST, so AI agents, apps and BI tools query governed models instead of writing raw SQL. It bundles agent skills, dashboards, materialization, and row-level access control.
Use Cases
Query governed data models from Claude, Cursor, Codex or custom agents over MCPGenerate Malloy queries from natural language instead of raw SQLDefine a semantic model once and serve it to agents, apps and BI toolsBuild Malloy-declared dashboards and no-build HTML data appsEnforce row-level access and authorization on agent queriesMaterialize and pre-aggregate expensive sources on demand or on a scheduleSearch and rank warehouse tables by plain-English descriptionRun analytical notebooks (.malloynb) against governed endpointsExpose an OpenAPI/REST query API for unattended agents and applicationsConnect local CSV, Parquet, JSON or Excel files via built-in DuckDBCompile and reload data model edits without restarting the serverCurate which sources are discoverable and queryable by agents
Built With
- Language
- TypeScript
- Frameworks
- Model Context Protocol · Malloy · DuckDB · OpenAPI · Bun · Node.js · Playwright · BigQuery · Snowflake · Postgres · Databricks · MotherDuck
Tags
mcp · mcp-server · semantic-modeling · text-to-sql · analytics · data-modeling · query-engine · dashboards · bi · ai-agents · row-level-access · materialization · duckdb · rest-api · data-apps · governance