I've seen three tools in the last few months claiming something similar. They all hallucinate less than GPT, but they still hallucinate. Is yours any different, if so then how?
The errata sheet and the datasheet are both ingested together and well in scope of Hydron,any work around to be done is verified by keeping both these docs in sight
That's a great question. Indexing time directly depends on the number of pages in the PDF, for a 5000 page PDF it roughly takes around 30 minutes to index.
Other files types supported include Word, Excel, CSV, Markdown, XML, Text and ZIP files. We can also index entire codebases and GitHub repositories as agent context.