Sources are where LD3 starts. Before any content is generated, you connect the places LD3 draws material from. A small set of trusted sources keeps fact extraction reliable.
What counts as a source
- RSS and Atom feeds* from blogs, news sites or industry publications.
- Newswires and CMS exports* you already publish with.
- Social inboxes* where briefs or mentions arrive.
- Briefing documents* you upload directly.
- Topic inputs* you want LD3 to monitor, which seed the same monitored source set.
How intake works
1. You connect a feed, site, inbox or document — or set a topic to monitor.
2. LD3 ingests each item with provenance metadata*: where it came from and when.
3. Every later draft can trace a claim back to its source. LD3 does not scrape anonymously; sources keep their original link and attribution.
What you control
Source intake is a configuration* step. LD3 runs the rest of the pipeline for you, but it cannot start until at least one source (or monitored topic) is connected.
Tips
- Prefer a few high-trust sources over many low-quality ones.
- Keep source links stable so verification and citations stay valid.
- Use Flows to refresh or rotate sources on a schedule.
Related
- [Set up an RSS source](/docs/ld3-rss-source-setup) — connect an RSS feed as a source.
- LD3 source verification — how sources are scored before use.
- LD3 topics — turning monitored sources into angles.
Was this article helpful?