Data Studio Agent. Your DBA.
Talk to your databases in plain language. Your agent understands your schema, runs the query, and walks you through what it did. Your 24/7 database engineer.
Get started in one command
codex mcp add data-studio -- npx -y @geek-fun/data-studio-mcpRequires DocKit and/or SqlKit installed, running, and connected to a database. The server auto-discovers running backends.
One MCP server, 70+ databases
Works with every major AI coding agent
Claude Code
Cursor
Windsurf
OpenCode
Codex
Cline
Pi
Qoder
Trae
Kiro
GitHub Copilot
CodeBuddy
Antigravity
Grok Build
Claude Code
Cursor
Windsurf
OpenCode
Codex
Cline
Pi
Qoder
Trae
Kiro
GitHub Copilot
CodeBuddy
Antigravity
Grok Build
OpenSearch
OpenSearchEvery database, every tool, one config
Ask your databases in plain language. Tell your agent what you need, it reads schemas, writes queries, and returns results. One MCP server config routes to the right backend across SQL and NoSQL, local and cloud, without per-database glue code.
Every tool your agent needs, one config
One MCP server exposes a focused toolset per backend. Your agent reads schemas, runs queries, and gets actionable diagnostics. No more, no less.
Core capabilities
One MCP server for your whole stack
A single npm package routes to both SqlKit (SQL) and DocKit (NoSQL) bridges over localhost. No per-database glue code. Register one MCP server, query everything.
Plain-language database work
No more writing every JOIN by hand or digging through CLI output. Ask for what you need and the agent executes verified queries against your real schemas.
Always-on diagnostics
Every tool reports backend availability and permission state. If a database app isn't running or an operation is gated, the agent gets an actionable message, not a silent failure.
Built on the apps you trust
All database drivers, SSH tunnels, and connection management live in DocKit and SqlKit. The MCP server is a thin routing layer that auto-discovers running backends.
Built for security-first teams
Credentials never leave the apps
The LLM only ever sees an opaque connection_id. Real credentials are resolved inside DocKit/SqlKit and never cross the MCP boundary. Your passwords and keys stay on your machine, in your app.
ID-based resource access
Agents access databases strictly by connection ID, never by embedding credentials in prompts or tool arguments. There is no path for the model to obtain or exfiltrate connection secrets.
Three-tier permission model
Read Only / Data Read-Write / Full Access modes gate every capability by risk level. Plus per-connection overrides: mark any connection read-only, or allowlist specific actions.
Explicit user confirmation
Destructive operations (DELETE, DROP, TRUNCATE) surface as Ask in the policy. The client prompts the user for explicit confirmation before anything destructive runs.
Action-level statement classification
SQL is parsed and classified by statement kind (Read / Write / Delete / DDL) before execution. Write-only tools reject DELETE statements, and delete tools reject DDL.
Local-only bridge
The bridge binds to 127.0.0.1 exclusively, unreachable from other machines. A thin routing layer with no server to host and no API keys to manage.
Permission modes
Give your coding agent access to your data
Install the MCP server, point it at DocKit and SqlKit, and start asking questions in plain language. Free, open source, Apache 2.0.
