A comment system backend on top of PostgreSQL
pg-discuss has a feature set comparable to popular commercial comment management systems. However, pg-discuss does not sell your users’ browsing habits and comments to marketers. You have full control over how your users’ information is used.
pg-discuss is a MIT-licensed open source project that allows you to operate, extend, and redistribute the application as you see fit. Contributions are welcome.
The pg-discuss API can be adapted to work with existing embeddable comment widgets. Of course, new widgets can also be written using the default REST API.
The application is highly extensible, with most default features implemented as extensions which are bundled. Extensions can change the comment workflow, comment metadata collection, rendering engine, authentication, and HTTP API format and endpoints. Extensions are implemented using a simple API that exposes logical plugpoints and events.
The core package of pg-discuss is less than 1000 lines. Code is structured for readability, conforms to PEP8, and has ample comments. It is written to be very approachable for both auditors and extension developers. Both prose and API documentation of internals are published.
PostgreSQL offers strong data guarantees, ease of management, and high performance. pg-discuss is written to take advantage of several PostgreSQL-specific features that enable flexible extension and low latency querying.
The pg-discuss application can be deployed across many hosts for redundancy or load balancing since it does not store any application state itself - all state is held in the PostgreSQL database or in the client browser session. pg-discuss also plays well with HTTP caches such as nginx, and a simple extension can inform the cache when to purge stale objects.
There are several good alternatives to pg-discuss that are also open source, but have somewhat different design goals.
Perhaps you don’t want to run a database?
Isso is an excellent SQLite-backed comment server - there is no separate database instance to run, since the data is stored in a file. Also written in Python. MIT licensed.