Contribute a source
RAG Repo gets better every time someone points us to a source we have missed. If you know a dataset, knowledge base, or repository that belongs here, we would love to add it.
Suggest a source
For now, the quickest way is to open an issue on GitHub. Tell us:
- the name of the source and a link to its official site
- what it contains, in a sentence or two
- why it is useful for RAG, and who might want it
- the licence, if you know it
That is all we need to get started. We will check the details, write it up in our house style, and add it to the directory.
What makes a good suggestion
We look for sources that are genuinely useful for retrieval, openly accessible or clearly worth the access they require, and documented well enough that we can describe them accurately. Primary sources only: link us to the people who publish the data, not a mirror or a reseller.
Coming soon
Direct community submissions, ratings, and collections are on the way. Before long you will be able to add and improve sources without leaving the site. Until then, GitHub issues are the way in, and every suggestion is read.