I watched your video - RAG App from Scratch: Verba 0.3.0
Are you guys planning for verba to ever be an enterprise-grade product?
Or is it more just a quick try-it-fast RAG app to get people excited about Weaviate?
I’m finding very few turnkey RAG apps available and Verba could be a big player here. But here’s what I think it would need (if helpful):
- Tenancy model - like some oppinionated way to add access control to the data.
- Helm chart (I know that weaviate has this) but would be cool to deploy verba into a cluster with lots of knobs turned on.
- More detailed information on performance and turning
- Crawlers - how can I ingest an entire corpus of s3 docs. Can verba do the ETL of a product like unstructured (or use it as a plugin) so that I’m not building my own preprocessing pipeline.
Looking forward to hearing your feedback on this.
Cheers.
hi @myriam.lachouch ! Welcome to our community
As of now, Verba is a demo on how to use Weaviate to do RAG.
But of course, as it is so cool, I also would love it to have more features and bring more value.
I believe it would be worth registering that in Verba GH GitHub - weaviate/Verba: Retrieval Augmented Generation (RAG) chatbot powered by Weaviate
Thanks!
1 Like
I used the AWS Command Line to sync down about 25k documents and had Verba ingest them locally.
1 Like
hi @mmike87 !!
Welcome to our community
Glad you are enjoying Verba!
Please, don’t forget to leave some starts in our github repos:
and
Thanks!
Hi Duda, I successfully implemented Verba-llama3-RAG solution locally. However, my repository has 100+ documents, so it is not practical to process one pdf at a time. Is there a way to bulk-upload all the documents for chunking and embeddings?
Hi @sk_perera !!
Welcome to our community
There will be some updates to Verba. While I can’t give ETA, I know that some features like this one are on the roadmap, as well as multi tenancy.
So this may be possible soon
Thanks!