SkillFoxx

pREST MCP

Instant REST and MCP over PostgreSQL: the agent reads an existing database through an MCP endpoint with no backend to write

  • 4,632 ★ on the repository
  • Last commit
  • MIT
MCP serverOfficial

Extended access

Why this level

  • Exposes PostgreSQL data through a network service
  • The full REST API can modify data with wrong permissions
All reasons and checks

prest/prestAuthor:pREST

Install

Manual install

docker run -e PREST_PG_URL=postgres://user:pass@host:5432/dbname -p 3000:3000 prest/prest

Runs pREST over your database. The connection can also be set via pg.* or DATABASE_URL. Enable MCP per the MCP over HTTP guide.

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • Exposes PostgreSQL data through a network service
  • The full REST API can modify data with wrong permissions

This is third-party code. Review the repository files before installing.

What it does

pREST stands up a ready API over an existing or new PostgreSQL database: table CRUD, custom SQL routes, auth, ACL and an MCP endpoint, all without hand-writing a backend. For agents there is MCP over HTTP that gives read access to the database, so an AI client can query tables and data through a standard protocol. Values in SQL templates can be passed by binding, so they travel to Postgres out of band from the query text and are never parsed as SQL. It installs via Docker, Homebrew or Go and points at the database through a connection variable.

Who it is for. For backend developers and data engineers who want fast PostgreSQL access through an API and MCP without their own service.

Good fit when

  • You want the agent to read PostgreSQL data over MCP
  • You want an instant REST API over an existing database
  • You want custom SQL routes with auth and ACL without a backend

Not a fit when

  • The database is not PostgreSQL
  • You cannot deploy and run a separate gateway service to the database

Example request

Through MCP show the orders table structure and select the last week's records

MCP

Transport
http
Authentication
not required

Environment variables

PREST_PG_URL required, secret
PostgreSQL connection string; pg.* or DATABASE_URL also work

Limitations

pREST is a separate service in front of PostgreSQL that you deploy and configure: connection, auth and ACL. The MCP endpoint gives read access, while the full REST API can change data, so set permissions and ACL for the agent's access. It needs PostgreSQL 9.5 or newer.

How to disable. Remove the pREST MCP endpoint address from the agent config and stop the pREST service.

FAQ

What does the agent get through MCP?

The MCP endpoint gives read access to database data through a standard protocol; changes go through the REST API with auth and ACL.

Which database is needed?

PostgreSQL 9.5 or newer, existing or new; pREST works on top of it.

Discussion

No questions or reviews yet.

MCP servers

A single Go binary turns a SQL database into a REST API with roles and an MCP server for Claude Code, Cursor and VS Code

MCP serverBroad access173

A PostgreSQL MCP server with index tuning, EXPLAIN plans, database health checks and a read-only mode

MCP serverBroad access3.5K
Official

A PostgreSQL backend your agent builds over MCP, where the platform verifies every change and can roll it back.

MCP serverExtended access83
Official

Two official Prisma MCP servers: a local one for migrations and Prisma Studio, and a remote one for managing Prisma Postgres databases

MCP serverBroad access49
Official

Supabase's official MCP server: the agent manages tables, reads project config and queries data after an OAuth login

MCP serverBroad access3K

README in short

The README describes pREST as a ready API that delivers REST and MCP over an existing or new PostgreSQL database: CRUD, custom SQL routes, auth, ACL and a read-only MCP endpoint. Install via Docker, Homebrew or Go, with the connection set through PREST_PG_URL or DATABASE_URL. It also covers safe value passing by binding rather than interpolation, MCP over HTTP, multi-database support and connecting AI clients such as Cursor and Claude. It includes Kubernetes deployment and one-click deploy. MIT licensed.

README badge

Report a problem