Weves

6 ms·

31.
by Weves 10mo ago
There are also many teams we work with that want to (1) retain model flexibility and (2) give everyone at the company the best model for the job. Every week? a model from a different provider comes out that is better at some tasks than anyo
32.
by Weves 10mo ago
Yea, the license is modeled after the Gitlab license. All of the core chat/RAG/agent logic is fully MIT, and >99% of deployments of Onyx are using the "community edition"!
33.
by Weves 10mo ago
You don't need any subscription to run the code! By default, none of the enterprise code runs (and it can all be completely removed and the app will work as expected). Fully FOSS version here: https://github.com/onyx-do
34.
by Weves 10mo ago
Broadly, I think other open source solutions are lacking in (1) integration of external knowledge into the chat (2) simple UX (3) complex "agent" flows. Both internal RAG and web search are hard to do well, and since we've st
35.
by Weves 10mo ago
Agree that's a lot of other projects out there, but why do you say the Vercel option is more advanced/mature? The common trend we've seen is that most of these other projects are okay for a true "just send messages to an
36.
by Weves 10mo ago
Some of the key differences: 1/ large connector suite + good RAG. Answers at scale is hard, and from our enterprise search roots, we've spent a lot of time with it. It's something that many teams expect from their chat UI. 2&
37.
by Weves 10mo ago
Thanks! Open source is awesome :)
38.
by Weves 10mo ago
RAG + connectors is a huge reason why people deploy Onyx (enterprise search roots means we do a pretty good job there). Also, open-source works really here, since connectors are a long-tail game. We've tried to make it easy to add conn
39.
by Weves 10mo ago
Haha, yea we've seen that exact story many times! Dissatisfied with Copilot and building a (not great) internal solution that is missing polish + most of the "advanced" feature set.
40.
by Weves 10mo ago
Really appreciate the feedback (and glad to hear the core concept resonated with you). The admin side of the house has been missing a bit of love, and we have a large overhaul coming soon that I'm hoping addresses some (most?) of your
41.
by Weves 10mo ago
"Chat UI" can "feel" a bit thin from an eng/product when you initially think about, and that's something we've had to grapple with over time. As we've dug deeper, my worry about that has gone down ove
42.
by Weves 10mo ago
Yes, it absolutely does! The chat UX, add-ons (deep research, code interpreter, RAG, etc.), and SSO are MIT licensed. Most deployments of Onyx are using the pure FOSS version of Onyx. Many individuals / teams have done extensive white
43.
Launch HN: Onyx (YC W24) – Open-source chat UI
254 points by Weves 10mo ago|160 comments
44.
Mouse could eavesdrop on you and rat you out(theregister.com)
6 points by Weves 1y ago|1 comments
45.
by Weves 1y ago
Don't mean to hijack (one of the Onyx founders here), but the example you described should be doable with Drive service accounts. Admittedly, our permissioning system is only implemented for a handful of connectors like Drive. Congratu
46.
Experts urge caution about using ChatGPT to pick stocks(arstechnica.com)
3 points by Weves 1y ago|0 comments
47.
Nintendo of America boss Doug Bowser is retiring(theverge.com)
3 points by Weves 1y ago|1 comments
48.
Elon Musk's Chatbot Goes to Washington(nytimes.com)
4 points by Weves 1y ago|0 comments
49.
by Weves 3y ago
We do have support for IdPs like Azure AD + role based access control (these aren't in the MIT version of Danswer though :sweat:). If you're interested in learning more, would love to chat through the details on Slack / Disco
50.
by Weves 3y ago
The name was intended to be (1) Danswer -> Deep (learning) Answer and (2) Danswer -> Dancer (thus the logo). Although we've heard quite a ton of different interpretations (the "dude named Dan" is quite popular, as is &q
51.
by Weves 3y ago
"We were building something similar in house, but then we found Danswer". That's something we keep hearing all the time :D (1) We handle chunking, embedding, building the keyword search index, etc. all on our end! We use Vesp
52.
by Weves 3y ago
Yep, this is exactly correct! For our SaaS version, we do have an agreement which allows us to look at data if needed to debug issues and/or improve search performance. For self-hosted deployments, usually a select few admins who have
53.
by Weves 3y ago
OpenAI's models may fit that description as well under the hood. Specifically, for `prefix-aware`, this is useful when you have short passages (e.g. Slack messages) that you are trying to match against short queries (e.g. user question
54.
by Weves 3y ago
Thank you for the kind words! Yea, the name has many meanings. It's a fun little puzzle to find them all :)
55.
by Weves 3y ago
Yes absolutely! We actually have a doc specifically for Ollama: https://docs.danswer.dev/gen_ai_configs/ollama
56.
by Weves 3y ago
Well the most standard approach is to use cross-encoders (e.g. something like Cohere Rerank) to give similarity scores between the query and the chunk, and then use these scores to update the ranking. Our approach is to use an LLM (gpt-3.5-
57.
by Weves 3y ago
So we're leaning on access control to do this! Right now we support manually configured group-based access at the connector level (e.g. Users X, Y, and Z make up the `Engineering` group and that group should have access to Folders A, B
58.
by Weves 3y ago
Chris here (the other founder) - yes you can! We have an `VALID_EMAIL_DOMAINS ` env variable which controls this. For example, for us we have `VALID_EMAIL_DOMAINS=danswer.ai`.
59.
by Weves 3y ago
Totally understand that :) With that said, we feel this is a completely different way of interacting with the system and therefore warranted a new post. Rather than just search, you can now converse with your docs while retaining all the sa
60.
Show HN: DanswerChat – open-source GPTs with access to your org's knowledge
14 points by Weves 3y ago|4 comments

More ›