Your AI Can Now See Your Databases
Claude, Cursor, and Copilot read live schemas, data, and streams via the built-in MCP server.
- 26 read-only tools
- No pasting schemas into chat
- Runs on your infrastructure
AI assistants are good at SQL and bad at your database — because they can't see it. So you become the courier: copy the DDL into chat, paste the error log, re-paste the schema after every migration. The answer you get is only as fresh as your last paste, and the moment something changes, the AI is reasoning about a database that no longer exists.
MCP (Model Context Protocol) removes the courier. Your AI client calls DBConvert Streams directly and works from the same live workspace you use — the connections, schemas, files, and stream state you already see in the app.
Real Prompts, Real Answers
Every exchange below came back from a live workspace — numbers and all, on the inexpensive Claude Haiku 4.5.
Stream troubleshooting
You
My MySQL → PostgreSQL CDC stream stopped overnight. List my streams, check status and recent errors, and tell me what to fix.
AI
Found it — the latest run of 10M-cdc-mysql2pg FAILED with Error 1049 (42000): Unknown database 'target', flagged source-side. The CDC source points at a database that doesn't exist on the MySQL server. Fix: create it (or correct the database name on the connection), then restart.
Analytics on live data
You
What are my top film categories by rental revenue? (sakila sample)
AI
Sports — $5,314.21 Sci-Fi — $4,756.98 Animation — $4,656.30 Drama — $4,587.39 Comedy — $4,383.58 Joined category → film_category → film → inventory → rental → payment, ran in ~55 ms. Want the query to reuse, or the full top 15?
Trimmed for the page — full transcripts: plain-English SQL and federated validation.
What Your AI Can Reach
You describe a goal in plain language; the AI picks and chains the tools itself.
Schema Inspection
- Describe any table or view — columns, keys, indexes
- Compare schemas between two databases
- Navigate databases, schemas, tables, and views
Read-Only SQL
- AI writes the query, runs it, hands back the rows
- EXPLAIN plans for index and performance review
- Preview table and view data on demand
Federated Queries
- JOIN PostgreSQL, MySQL, and Parquet in one SELECT
- No ETL, no staging table — data stays where it lives
- EXPLAIN federated plans before running them
Stream Diagnostics
- Status, recent errors, and logs per stream
- Throughput stats and run history
- AI pinpoints the failure; you apply the fix
Files & S3
- Browse local files and S3 buckets and objects
- Read CSV, Parquet, JSON, and JSONL schemas
- Query file contents with the same SQL tools
Cross-Source Validation
- Diff a file export against the live table
- Compare data samples across two sources
- Row counts matched in a single federated query
Plain-language map of all 26 tools in the tools reference .
Read-Only by Design
The AI can look and advise. It cannot change anything.
SELECT-Only Filter
Only SELECT and WITH … SELECT pass the server-side safety filter. INSERT, UPDATE, DELETE, DROP, ALTER, and CREATE are rejected at the server before they reach your database.
No Credential Access
Connections are read from your workspace; the AI never sees passwords and cannot create or edit connections. Secrets are redacted from everything it reads.
No Stream Control
The AI can't start, stop, pause, or reset streams. It surfaces the status, the error, and the fix — you (or the UI) run the action.
Your Infrastructure Only
The server runs on your machine or in your stack and only answers your client's requests. It never calls out to DBConvert or your AI provider on its own.
A fast analyst on tap — not an autopilot for your production database.
Full boundary in Safety & privacy.
Always Live, Never Snapshotted
Edit a connection in the DBConvert Streams UI and the MCP server picks up the change in a fraction of a second — no restart of your AI client. The AI reads through the same workspace that powers Data Explorer , so it sees exactly what you see: no separate credentials, no second copy of your schema to keep in sync.
Two Ways to Connect
Pick by how DBConvert Streams runs. Step-by-step guides for every client:
- Cursor
- VS Code + GitHub Copilot
- Windsurf
- Claude Code
- Claude Desktop
- Gemini CLI
- Codex CLI
Desktop App → stdio
Open the ✨ AI Assistants panel, click Connect next to your client, restart it. The panel writes the client config for you.
Docker / Remote → HTTP
Point your client at http(s)://<host>/mcp with your API key — works for a local container or a server across the network.
Connect Your AI Assistant in Minutes
Then ask it to list your connections — if they come back, you're done.
Prefer scripted automation over conversation? See the REST API.