Discourse-reader – a typed, read-only Python client for Discourse

Hi HN,

I've had to pull data from Discourse forums a few times now, and always had to keep dealing with raw requests.get(), rate-limiting and navigating the returned JSON. Our ecosystem is in Python so I looked if there was already a solution, I found pydiscourse (https://github.com/pydiscourse/pydiscourse), but it requires setting up an API key (a hassle) and only returns raw dicts.

There were good solutions in other languages (like discourse2 in TypeScript) but yeah that wasn't in Python.

This is useful if you don't want to set up an API key, only care about retrieving data (not interacting with the platform) and want something in Python. Also has some basic rate-limiting built in!

Example usage:

from discourse_reader import DiscourseClient client = DiscourseClient("https://meta.discourse.org") topic = client.topics.get(12345) print(topic.posts) print(topic.accepted_answer.cooked)

for topic in client.topics.latest(limit=100): print(f"{topic.views:>6} views {topic.title}")

The Pydantic models are set with extra="allow", so if you use this to hit forums with plugins that have extra fields those won't be dropped.