github.devotd.
NewData

postgrest/postgrest v16.5 — Exposing PostgreSQL Schemas as REST APIs

PostgREST v16.5 exposes the schema of a PostgreSQL database as a REST API, and no release notes are available for this version.

Repositorypostgrest/postgrest ↗

PostgREST v16.5 exposes the schema of a PostgreSQL database as a REST API, and no release notes are available for this version.

GitHub description: REST API for any Postgres database

stars
★ 27.7K
Forks
1.2K
Language
Haskell
License
MIT
Latest release
v16.5
Read original source ↗View on GitHub ↗

The latest release of the GitHub repository postgrest/postgrest is v16.5. PostgREST is a standalone server that reads the schema of an existing PostgreSQL database and exposes it as a REST API. It builds the service layer from the database alone, without writing a separate CRUD server for relational data. Because it only needs a Postgres database, it is not tied to any particular backend. What changed in v16.5 cannot be determined, because the given material contains no release notes.

The README states that PostgREST provides a complete RESTful API from an existing PostgreSQL database. The implementation is built with Haskell and Warp, and it leaves JSON serialization (the process of converting data into JSON format), data validation, and authorization (deciding who may access which data) to the database as much as possible. It also supports JWT authentication for verifying identity with tokens, managing API versions as database schemas, and self-documenting output generated from the OpenAPI specification. The implementation language is Haskell, the license is MIT, the repository was started on June 13, 2014, and the star count is recorded as 27726.

However, the momentum score in the research record is low at 1.9 out of 10, and the freshness, technology, and quality scores are not recorded. The given information contains no performance figures showing request throughput or response time, and no comparison with other API servers. There is also no material describing behavior in environments with many concurrent requests or large data volumes. Without performance evidence, it is hard to decide on adoption right away. It remains a project worth checking for teams that already use a PostgreSQL database and need somewhere to put read and write APIs.