{"id":"pgmcp","name":"PGMCP","homepage":"https://github.com/subnetmarco/pgmcp","repo_url":"https://github.com/subnetmarco/pgmcp","category":"database","subcategories":["natural-language-query","mcp-server","postgresql"],"tags":["postgres","sql","natural-language","mcp","read-only","streaming"],"what_it_does":"MCP server that converts natural language questions into SQL queries against any PostgreSQL database, returning results in table, JSON, or CSV format with streaming support for large result sets.","use_cases":["Ask plain-English questions against a production Postgres database without writing SQL","Export database data to CSV/JSON via natural language commands","Schema discovery and exploration through conversational queries","Business analytics queries from non-technical users via AI assistants","Read-only database auditing and reporting workflows"],"not_for":["Write operations — enforces read-only access by design","Non-PostgreSQL databases","Environments where connecting an AI to production DB credentials is not acceptable"],"best_when":"You want a safe, read-only natural language interface to an existing Postgres database and are using a MCP-compatible client like Claude Desktop or Cursor.","avoid_when":"You need write access, multi-database federation, or are working with databases other than PostgreSQL.","alternatives":["pgvecto.rs","supabase MCP","datasette","text2sql custom chains"],"af_score":73.3,"security_score":72.0,"reliability_score":null,"package_type":"mcp_server","discovery_source":["github"],"priority":"low","status":"evaluated","version_evaluated":"latest","last_evaluated":"2026-03-01T09:50:06.073923+00:00","performance":{"latency_p50_ms":null,"latency_p99_ms":null,"uptime_sla_percent":null,"rate_limits":null,"data_source":"llm_estimated","measured_on":null}}