{"id":"cdatasoftware-postgresql-mcp-server-by-cdata","name":"postgresql-mcp-server-by-cdata","homepage":null,"repo_url":"https://github.com/CDataSoftware/postgresql-mcp-server-by-cdata","category":"devtools","subcategories":[],"tags":["mcp","postgresql","jdbc","claude-desktop","stdio","read-only","local-integration"],"what_it_does":"Provides a local MCP (Model Context Protocol) server that exposes data from a PostgreSQL database via CData JDBC Driver as MCP tools, primarily for reading (tables/columns) and running SELECT queries over stdio to an MCP client like Claude Desktop. It is configured by building a JAR and providing a JDBC connection .prp file.","use_cases":["Letting an LLM query live PostgreSQL data via natural-language questions without writing SQL","Discovering available tables and columns in a connected PostgreSQL data source","Executing read-only analytical queries (SELECT) through an MCP client"],"not_for":["Direct use as a networked API service (the server is stdio-only and local-machine bound)","Production-grade CRUD workflows as described (the local MCP server is explicitly read-only in this project)","Use cases requiring fine-grained query safety controls (tooling exposes arbitrary SELECT SQL input)"],"best_when":"You need local, read-oriented access from an MCP-capable desktop client to PostgreSQL data using existing CData JDBC connectivity.","avoid_when":"You need remote/multi-tenant access, strong guardrails around what queries can be executed, or an HTTP/REST interface with standard operational tooling.","alternatives":["CData managed MCP platform (CData Connect AI) referenced by the README","A custom MCP server or direct integration that exposes a constrained query interface","Use PostgreSQL directly with a query tool via your own application/API layer (with query validation)"],"af_score":50.2,"security_score":33.2,"reliability_score":23.8,"package_type":"mcp_server","discovery_source":["github"],"priority":"low","status":"evaluated","version_evaluated":null,"last_evaluated":"2026-04-04T20:00:02.151659+00:00","interface":{"has_rest_api":false,"has_graphql":false,"has_grpc":false,"has_mcp_server":true,"mcp_server_url":null,"has_sdk":false,"sdk_languages":[],"openapi_spec_url":null,"webhooks":false},"auth":{"methods":["JDBC connection authentication via CData JDBC Driver (configured in .prp / JDBC Url)"],"oauth":false,"scopes":false,"notes":"Authentication is delegated to the underlying CData JDBC Driver configuration included in the .prp/JdbcUrl. README mentions OAuth may be required for some data sources, but no MCP-level OAuth/auth flow is described."},"pricing":{"model":null,"free_tier_exists":false,"free_tier_limits":null,"paid_tiers":[],"requires_credit_card":false,"estimated_workload_costs":null,"notes":"No pricing for the MCP server itself is provided in the README; CData JDBC Driver licensing is required (the README includes a driver license step)."},"requirements":{"requires_signup":false,"requires_credit_card":false,"domain_verification":false,"data_residency":[],"compliance":[],"min_contract":null},"agent_readiness":{"af_score":50.2,"security_score":33.2,"reliability_score":23.8,"mcp_server_quality":70.0,"documentation_accuracy":65.0,"error_message_quality":0.0,"error_message_notes":null,"auth_complexity":55.0,"rate_limit_clarity":0.0,"tls_enforcement":40.0,"auth_strength":45.0,"scope_granularity":10.0,"dependency_hygiene":40.0,"secret_handling":30.0,"security_notes":"Security controls are largely inherited from the CData JDBC Driver and the database connection. The README does not describe MCP-level access control, TLS requirements for the database connection, secrets storage hygiene, query validation/sanitization, or protection against prompt-driven expensive/broad SELECTs. Since it accepts SQL input for run_query, guardrails are important at the client/agent layer.","uptime_documented":0.0,"version_stability":45.0,"breaking_changes_history":30.0,"error_recovery":20.0,"idempotency_support":"false","idempotency_notes":null,"pagination_style":"none","retry_guidance_documented":false,"known_agent_gotchas":["Server communicates via stdio; it can only be used with clients on the same machine","Tool outputs for tables/columns are returned in CSV format; agents may need to parse CSV carefully","The run_query tool accepts arbitrary SQL (even though the project is marketed read-only), so agents can generate expensive or broad queries unless the client/agent constrains it"]}}