FreemiumVisit Website
Pros & Cons
Pros
- Open-source core with Apache 2.0 license
- Prevents AI hallucinations with governed metrics
- Multi-interface: React, MCP, REST, GraphQL
- Built-in multi-tenancy and row-level security
Cons
- Early-stage product (founded 2025)
- Cloud pricing not publicly listed
- Limited to B2B analytics use cases
- Requires data warehouse setup
Key Features
Semantic metric layer (YAML)Row-level security per tenantMCP server for AI agentsReact SDK for embedded chartsREST & GraphQL APIsDBT model auto-detectionAudit loggingSelf-hosted or cloud deployment
Pricing Plans
7-day Free TrialOpen Source
Free
- Apache 2.0 license
- Self-hosted via Docker Compose
- CLI workflow
- Core semantic layer
Cloud
Usage-based
- Managed infrastructure
- Scale daily API calls
- 7-day free trial
- Audit logging
About Bonnard
By Toolradar Team·
Bonnard is an agentic analytics layer that enables B2B SaaS companies to embed governed AI analytics into their products. Define metrics once in YAML, deploy them across React charts, MCP servers, REST and GraphQL APIs with built-in row-level security and multi-tenancy. Connects to data warehouses with automatic DBT model detection and prevents hallucinated aggregations by having AI agents query governed metric definitions.
Reviews
No reviews yet. Be the first to review Bonnard!
Explore More
Bonnard FAQ
How does Bonnard prevent AI hallucinations in analytics?
Bonnard uses a governed semantic layer where metrics are pre-defined in YAML. AI agents query these governed definitions rather than raw database tables, ensuring accurate aggregations and preventing hallucinated results.
Can I self-host Bonnard?
Yes, Bonnard offers an open-source Apache 2.0 licensed version that can be self-hosted via Docker Compose. There is also a managed cloud option for teams that prefer not to manage infrastructure.
What data warehouses does Bonnard support?
Bonnard connects to standard data warehouses and automatically detects DBT models. It creates views and cubes from raw tables with built-in multi-tenant isolation.
How does the MCP server integration work?
Bonnard includes an MCP server that exposes your governed metrics to AI agents. Agents can query defined metrics through the Model Context Protocol, ensuring they always use accurate, governed data definitions.
What security features does Bonnard provide for multi-tenant analytics?
Bonnard enforces row-level security per tenant with token-based authentication mapping. Every query automatically respects tenant boundaries, and all access is logged for compliance auditing.
Source: bonnard.dev