MCP for AI Agents

MCP for AI Agents

pgvis automatically generates Model Context Protocol tools from your database schema. Every table becomes a set of typed tools that LLM agents can discover and call.

What Gets Generated

For each table in your exposed schemas, pgvis creates:

Tool	Purpose
list_{table}	Query rows with filtering, ordering, pagination
create_{table}	Insert one or more rows
update_{table}	Update rows matching a filter
delete_{table}	Delete rows matching a filter

For each function:

Tool	Purpose
call_{function}	Call the database function with typed arguments

Stdio Transport (Claude Desktop)

Add to your Claude Desktop MCP configuration:

{
  "mcpServers": {
    "pgvis": {
      "command": "pgvis",
      "args": ["--dsn", "postgres://user@localhost/mydb", "mcp"]
    }
  }
}

HTTP Transport

Run alongside the REST API:

pgvis --dsn "postgres://user@localhost/mydb" serve --mcp-http

The MCP endpoint is available at /mcp.

Discovery Resources

pgvis exposes MCP resources for schema discovery:

• pgvis://schemas — list all available schemas
• pgvis://{schema}/schema — describe tables, columns, and relationships in a schema

Tool Input Schema

Each tool has a typed JSON Schema for its inputs. For example, list_users might accept:

{
  "select": "id,name,email",
  "filter": { "active": "eq.true" },
  "order": "name.asc",
  "limit": 10
}

Embedding in Your Own App

use pgvis_lib::Builder;

let mcp = Builder::new("postgres://localhost/mydb")
    .schemas(vec!["public"])
    .build_mcp_server()
    .await?;

pgvis_lib::pgvis_mcp::serve_stdio(mcp).await?;