Provides an open-source, self-hosted semantic layer for AI agents, ensuring reliable and governed data outputs.
Sponsored
Bonnard is designed as an agent-native semantic layer, offering a unified source of truth for metric definitions across all data consumers, including AI agents, applications, and dashboards. It tackles the challenge of inconsistent data outputs by providing a self-hosted Docker deployment, allowing users to run Bonnard on their own infrastructure without requiring a cloud account. This robust system facilitates seamless integration with various AI agents via the Model Context Protocol (MCP), enabling them to query data reliably from your defined semantic layer.
Key Features
01Support for multiple data warehouses including Snowflake, BigQuery, PostgreSQL, and DuckDB
02Deploy API for pushing schema updates without server restarts
03SQL-based Cube semantic layer with caching, access control, and multi-database support
04MCP server for AI agents to query the semantic layer
05Admin UI to browse deployed models, views, and measures
068 GitHub stars
Use Cases
01Deploying a self-hosted, agent-native semantic layer for AI agents
02Enabling AI agents to reliably query structured data and build data-driven applications
03Standardizing metric definitions and ensuring governed data outputs across all data consumers
