VeexaDB Live Read the docs
Database

The multi-tenant database you run, with an AI-native API.

VeexaDB is a self-hosted, multi-tenant document and table database. Reach it over a language-agnostic REST API or a native MCP interface — with RBAC, indexes, and vector and full-text search — and keep your data on your own infrastructure. No driver to install, no schema migration to run first.

Why VeexaDB

Multi-tenant, self-hosted, and built for AI.

Multi-tenant by design

Per-tenant isolation with project- and environment-level scoping and RBAC built in — not bolted on at the app layer.

REST and MCP

A language-agnostic HTTP API with no driver to install, plus a native MCP interface — no SDK lock-in.

AI-native

The MCP interface, and the db-agent multi-agent layer, let assistants read, write, and reason over your data directly.

Vector and full-text search

Semantic vector search and full-text search are built in, so VeexaDB doubles as a retrieval store for RAG.

Runs on your infrastructure

Self-hosted on the Veeexa platform — your data and your database stay under your control.

Documents and tables

Model data as documents or tables — whichever fits your app — with indexes to keep lookups fast.

All features →

One platform

Your data, your infrastructure — wired into Veeexa.

Your infrastructureVeexaDB is self-hosted on the Veeexa platform, so your data and your database stay under your control.
One identityVXAuth signs into VeexaDB and the rest of the platform; the MCP interface auto-registers with it.
AI-nativeThe MCP interface and the db-agent multi-agent layer let assistants read, write, and reason over your data.
Works with VXS3Keep large objects in object storage and their metadata in VeexaDB, on one Veeexa bill.