RAG to is not just context window, it's using tools like qdrant.
I'm asking because I'm trying to figure out if you have your own transcripts that your creating ofr this or if you're somehow using pu...
That is incorrect. Vector lookup like qdrant is just one technique that classifies as RAG. RAG is any technique that feeds external information to an LLM's context window. The technique can be keyword search, graph db traversal, vector search, web search or even SQL db queries.
In our case the endpoint has access to our vast hybrid (keyword and vector) database or 150k+ hours and runs an agentic loop several times to find you answers at a flat rate of $0.10. I wrote up an explanation here: https://primal.net/e/nevent1qqsdem4hd9tp67g4hwjddk46009scasy6fxmf0esdvd0k5sz9uwy38g5kd44m also note correction here: https://primal.net/e/nevent1qqsvxkey78ntdth49800ew2e383u2m30gc2d7gj4p2uc9wnag7huttc50p46x
We source all the transcripts so you dont have to. To do it on your own youd conservatively spend like $30k just for the transcripts let alone the hosting costs. Just use Jamie :)
In our case the endpoint has access to our vast hybrid (keyword and vector) database or 150k+ hours and runs an agentic loop several times to find you answers at a flat rate of $0.10. I wrote up an explanation here: https://primal.net/e/nevent1qqsdem4hd9tp67g4hwjddk46009scasy6fxmf0esdvd0k5sz9uwy38g5kd44m also note correction here: https://primal.net/e/nevent1qqsvxkey78ntdth49800ew2e383u2m30gc2d7gj4p2uc9wnag7huttc50p46x
We source all the transcripts so you dont have to. To do it on your own youd conservatively spend like $30k just for the transcripts let alone the hosting costs. Just use Jamie :)