DataMCPDataMCP

Roadmap

We're building the layer between your data and AI - starting with external tools, moving to in-dashboard chat, then private networks.

Q2 2026

PostgreSQL MCP server

Done

Core MCP server with schema access, SQL queries, and fine-grained permissions. Works with Cursor, Claude, VS Code, Windsurf, and any MCP client.

AI schema descriptions

Done

Auto-generated table and column descriptions so AI tools understand your data structure without extra prompting.

Connection pooling

Done

Persistent connection pools for fast, reliable query execution across concurrent AI sessions.

ChatGPT integration

Done

Native OAuth connection for ChatGPT - query your PostgreSQL database in plain English directly from chat, no SQL required.

Q3 2026

In-dashboard chat

Planned

Chat with your database directly inside DataMCP - no external AI tool needed. The fastest way to go from connection string to your first answer.

DataMCP Agent

Planned

A lightweight agent that runs inside your private VPC and connects outbound to DataMCP. Full product experience with zero inbound firewall changes - no self-hosting required.

Column-level masking

Planned

Mask PII columns (email, phone, SSN) in query results per MCP link. Ship AI features without exposing sensitive data to external models.

MySQL support

Planned

Connect MySQL databases with the same permission layer, audit trail, and MCP tools as PostgreSQL.

Q4 2026

SSO / SAML

Exploring

Enterprise single sign-on for teams using Okta, Azure AD, or Google Workspace.

Multi-database queries

Exploring

Join and query data across multiple connected databases in a single MCP call.

Audit export

Exploring

Export query logs as structured reports for SOC 2, HIPAA, and internal compliance audits.

2027

Self-hosted edition

Exploring

Full on-premise deployment for air-gapped environments and teams with strict data residency requirements.