Inconvo

Inconvo is an open-source platform for building customer-facing data agents that let users query production SQL databases through natural language. Backed by Y Combinator, it handles the hard parts of data agent development — safe query validation, multi-tenancy, permissions, observability, and conversation state — so developers can ship reliable chat-with-data features in minutes. Inconvo integrates via API/SDK or as an MCP server, and supports databases including PostgreSQL, MySQL, Microsoft SQL Server, Amazon Redshift, BigQuery, and Supabase.

• Semantic Layer — Define how agents interpret your data, enforce permissions, apply business rules, and scope access per tenant without writing custom middleware.
• Safe Query Execution — Validates every query before converting to SQL, ensuring queries are always safe, permissioned, and never destructive.
• Multi-Tenancy Support — Built-in tenant scoping lets you define data access at a per-tenant level, making it production-ready for SaaS products.
• API & SDK Integration — Integrate agents into your existing codebase using the Node.js SDK or REST API with response streaming support.
• MCP Server Deployment — Deploy agents as MCP servers on your own domain with built-in OAuth support in a few clicks, compatible with ChatGPT, Claude, and Microsoft Copilot.
• Observability & Agent Traces — Full visibility into every decision an agent makes, with conversation logs, usage analytics, and step-by-step traces for debugging.
• Conversation State Management — Manages multi-turn conversation context automatically, enabling contextual follow-up questions without extra engineering.
• Database Connections — Supports PostgreSQL, MySQL, MsSQL, Amazon Redshift, BigQuery, and Supabase out of the box.
• CLI Quickstart — Run npx inconvo@latest dev to start locally and get your first data agent running in under five minutes.
• Assistant-UI & Vercel AI SDK — Pre-built integrations with popular frontend frameworks for embedding chat interfaces directly into your product.