datatape

PostgreSQL MCP server

Give AI clients reviewed tools for PostgreSQL

Expose fixed, parameterized PostgreSQL queries as MCP tools from a hosted endpoint, with a practical read-only operating model.

Published
2026-08-31
Reviewed by
Rachid

Direct answer

Direct answer

Use Datatape with PostgreSQL when the workflow is known in advance and should be shared safely: define the query, expose only the parameters an agent needs, test the result, and publish the tool. For ad hoc schema exploration, use a database-native client instead.

PostgreSQL
Scoped role
Reviewed tool
Typed inputs
MCP endpoint
Authenticated
AI client
Stable contract
The MCP layer narrows the interface. Database permissions remain the enforcement boundary.
Product evidence · current source form
Datatape form for adding a PostgreSQL source

Best for repeated operational questions

A support assistant may need customer health, an incident summary, or a billing status. Those are stable questions with known inputs. Encoding each as a tool keeps the query logic out of the prompt and makes the interface understandable to both operators and clients.

When to choose another path

If an analyst needs to discover tables, compose novel joins, or iterate freely, a SQL editor or tightly controlled text-to-SQL environment is a better fit. Datatape’s strength is a reviewed contract for recurring work.

Example contract

customer_health

Return the current health signals for one customer account.

account_id
string · required
Reviewed query
SELECT account_id, plan, last_seen_at, open_incidents
FROM customer_health
WHERE account_id = {{ account_id }}

Returns: One account-level row with the fields the client needs to answer the workflow.

A practical operating model

MCP does not replace database security. Pair the tool contract with controls at the source.

Start with a read-only role

Create a dedicated PostgreSQL login with access only to the schemas and views these tools require. Database grants—not prompt instructions—are the enforcement boundary.

Keep parameters typed and narrow

Expose identifiers and bounded filters instead of accepting arbitrary SQL fragments. Test representative and empty results before publishing.

Review changes as interface changes

Treat a query or parameter edit like an API change. Re-test downstream expectations before making the revised tool available.

Client setup

Connect the clients your team already uses

The same hosted endpoint can be registered in multiple compatible clients. OAuth is the default human flow; agent API keys support non-interactive clients where headers are available.

Add the remote server from Settings → Connectors, then complete the browser authorization flow.

  1. Step 1Open Settings → Connectors and choose Add custom connector.
  2. Step 2Enter https://<org-id>.mcp.datatape.ai/mcp as the remote MCP server URL.
  3. Step 3Connect and finish authorization in the browser.

Client settings can change between releases. Verify the transport and authentication flow against the current client documentation.

Evidence

Sources and review

Written by Datatape Editorial. Product and technical claims reviewed by Rachid.

Continue