Datris.ai has open sourced what it calls the first MCP-native data platform, giving AI agents direct control over the full data lifecycle through 30+ native tools while eliminating vendor lock-in with a fully deployable open stack.
Datris.ai has announced the open-source release of Datris Platform, positioning it as the first data platform built with Model Context Protocol (MCP) as its primary interface. The platform has been designed natively for AI agents, enabling them to function as first-class pipeline operators instead of being added as an afterthought to legacy data systems.
At the core of the release is a native MCP server with more than 30 tools spanning pipeline creation, ingestion, validation, transformation, querying, vector search, RAG workflows, profiling, metadata discovery, and error diagnosis. The platform works with any MCP-compatible agent, including Claude, Cursor, and custom frameworks, allowing autonomous control of the entire data lifecycle without custom integration code.
“We built the MCP server first and made everything accessible through it,” said Todd Fearn, Founder and CEO of Datris.ai. “Most platforms bolt on an AI layer after the fact. We wanted agents to be native operators, not second-class citizens.”
The full stack is built on open technologies including Spring Boot, Docker Compose, MinIO, MongoDB, PostgreSQL, ActiveMQ, and HashiCorp Vault, with support for Anthropic Claude, OpenAI, and fully local, air-gapped deployments via Ollama. A single Docker Compose deployment brings up the full environment, reinforcing the platform’s open-source, no-lock-in positioning.
Datris.ai has made the platform, documentation, and GitHub repository available immediately, alongside Market Macro Agent, a real-time financial data AI agent use case.
