@supabase/mcp-server-postgrest
MCP ServerFreeMCP server for PostgREST
Capabilities5 decomposed
postgrest api schema introspection and tool registration
Medium confidenceAutomatically discovers PostgREST-exposed database tables, views, and functions by querying the OpenAPI schema endpoint, then registers them as MCP tools with typed parameters derived from the schema. Uses the MCP tool registry pattern to expose database operations as callable functions with JSON Schema validation, enabling LLM agents to understand available database operations without manual configuration.
Implements dynamic schema-to-MCP-tool mapping by parsing PostgREST OpenAPI schemas, eliminating manual tool definition — the server introspects the database API surface and auto-generates typed MCP tools with parameter validation derived directly from the REST schema
Tighter integration with PostgREST than generic REST-to-MCP adapters because it understands PostgREST-specific conventions (RPC endpoints, table naming, auth headers) and automatically maps them to safe, typed tool calls
parameterized database query execution via mcp tool calls
Medium confidenceTranslates MCP tool invocations into HTTP requests to PostgREST endpoints, handling parameter binding, query string construction, and response parsing. Supports filtering, sorting, pagination, and RPC function calls by mapping tool arguments to PostgREST query parameters and request bodies, with automatic JSON response deserialization and error handling.
Implements a parameter-to-PostgREST-query-syntax compiler that translates typed MCP tool arguments into correct HTTP query strings and request bodies, handling PostgREST-specific operators (eq, gt, in, etc.) and response formats without requiring the LLM to understand REST conventions
More ergonomic than raw REST clients because it abstracts PostgREST's query syntax into typed function parameters, and more flexible than SQL-only approaches because it supports both queries and mutations through a unified tool interface
authentication header injection and credential management
Medium confidenceManages PostgREST authentication by injecting API keys or JWT tokens into HTTP request headers for each tool invocation. Supports both Supabase API key authentication and custom JWT tokens, with configurable header names and token refresh logic to maintain valid credentials across multiple database operations.
Implements a credential injection layer that sits between the MCP tool interface and PostgREST HTTP calls, allowing the server to act as a trusted intermediary that holds credentials and enforces authentication without exposing keys to the LLM client
Safer than passing credentials to the LLM client because authentication is handled server-side, and more flexible than hardcoded credentials because it supports both API keys and JWT tokens with configurable header injection
mcp server lifecycle and tool registry management
Medium confidenceImplements the MCP server protocol by managing tool registration, request routing, and response serialization. Handles MCP message parsing, tool discovery requests, tool call invocations, and error responses according to the MCP specification, with proper async/await patterns for non-blocking database operations.
Implements the full MCP server protocol stack for PostgREST, handling tool discovery, invocation routing, and response serialization as a thin adapter layer that translates MCP calls to PostgREST HTTP requests
Purpose-built for PostgREST integration rather than a generic MCP framework, making it simpler to deploy but less flexible for non-PostgREST use cases
error handling and response normalization
Medium confidenceCatches PostgREST HTTP errors (4xx, 5xx responses) and normalizes them into MCP-compatible error responses with descriptive messages. Handles common database errors (constraint violations, authentication failures, not found) and translates them into human-readable error messages that LLM clients can understand and act upon.
Translates PostgREST HTTP error responses into MCP-compatible error formats with contextual messages, allowing LLM clients to understand database failures without parsing raw HTTP status codes
More user-friendly than raw HTTP errors because it provides semantic error messages that LLMs can understand and act upon, improving agent reliability
Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.
Related Artifactssharing capabilities
Artifacts that share capabilities with @supabase/mcp-server-postgrest, ranked by overlap. Discovered automatically through the match graph.
@bunli/plugin-mcp
MCP (Model Context Protocol) plugin for Bunli - create CLI commands from MCP tool schemas
@maz-ui/mcp
Maz-UI ModelContextProtocol Client
mcp-security-hub
A growing collection of MCP servers bringing offensive security tools to AI assistants. Nmap, Ghidra, Nuclei, SQLMap, Hashcat and more.
Hippycampus
** - Turns any Swagger/OpenAPI REST endpoint with a yaml/json definition into an MCP Server with Langchain/Langflow integration automatically.
Supabase
** - Connects to Supabase platform for database, auth, edge functions and more.
user-postgresql-mcp
A PostgreSQL MCP server built with @modelcontextprotocol/sdk.
Best For
- ✓Teams building LLM agents that need database read/write access via PostgREST
- ✓Developers integrating Supabase databases with Claude or other MCP-compatible LLM clients
- ✓Builders prototyping AI applications that require dynamic schema discovery from existing REST APIs
- ✓LLM agents that need to query or mutate database state through natural language
- ✓Applications where the AI should execute database operations without direct SQL access
- ✓Teams using Supabase who want Claude or other LLMs to interact with their database safely
- ✓Production deployments where database credentials must not be exposed to LLM clients
- ✓Multi-tenant applications using Supabase RLS policies that require per-user JWT tokens
Known Limitations
- ⚠Requires PostgREST to expose OpenAPI schema at /rpc endpoint — custom PostgREST deployments may need configuration
- ⚠Tool registration is static at server startup — schema changes require server restart to reflect new tables/functions
- ⚠No built-in caching of schema introspection — each server start re-fetches the full OpenAPI schema
- ⚠Limited to operations exposed by PostgREST (SELECT, INSERT, UPDATE, DELETE, RPC calls) — cannot access database features not surfaced by the REST layer
- ⚠No transaction support — each tool call is a separate HTTP request with no multi-statement atomicity
- ⚠Filtering capabilities limited to PostgREST's query syntax (eq, gt, lt, in, etc.) — complex WHERE clauses must be expressed as multiple filter parameters
Requirements
Input / Output
UnfragileRank
UnfragileRank is computed from adoption signals, documentation quality, ecosystem connectivity, match graph feedback, and freshness. No artifact can pay for a higher rank.
Package Details
About
MCP server for PostgREST
Categories
Alternatives to @supabase/mcp-server-postgrest
Are you the builder of @supabase/mcp-server-postgrest?
Claim this artifact to get a verified badge, access match analytics, see which intents users search for, and manage your listing.
Get the weekly brief
New tools, rising stars, and what's actually worth your time. No spam.
Data Sources
Looking for something else?
Search →