sourceharbor

<p align="center"> </p>

<p align="center"> <strong>Turn YouTube, Bilibili, and RSS sources into readable digests, traceable job runs, and searchable knowledge for operators and agents.</strong> </p>

<p align="center"> <a href="#see-it-in-30-seconds">See It In 30 Seconds</a> · <a href="./docs/start-here.md">Quickstart</a> · <a href="./docs/see-it-fast.md">No-Boot Tour</a> · <a href="./docs/index.md">Docs Home</a> · <a href="./docs/proof.md">Proof</a> · <a href="./docs/compare.md">Why It Stands Out</a> · <a href="https://github.com/xiaojiou176-open/sourceharbor/discussions">Discussions</a> </p>

<p align="center"> </p>

SourceHarbor is a GitHub-native knowledge pipeline for teams who need more than a transcript dump.

It helps you subscribe to long-form sources, pull fresh items, turn them into digests, inspect every step, ship artifacts, and reuse the same knowledge surface through API, MCP, and a browser command center.

If you only have half a minute, do not start with setup. Start with the three surfaces that explain the product fastest:

1. Command center: one operator view for subscriptions, intake, job counts, and recent artifacts.
2. Digest feed: a reading flow where entries such as AI Weekly and Digest One become reusable summaries instead of lost li