Open Source Unified Search across Company Tools

Danswer is the AI assistant that connects to all your company tools and docs and makes finding any piece of information a breeze. Think ChatGPT, but with access to all your companies documents (only the ones you have access to, of course :P). Or, if you know Glean, you can think about it like an open-source Glean. Concerned about your data? Don't worry we're fully open source (MIT license) and built to be self-hosted. You can setup Danswer in your VPC in less than 30 minutes - no packets ever have to leave your cloud. Check us out on Github: https://github.com/danswer-ai/danswer!

Team Size:2
Location:San Francisco
Group Partner:Jared Friedman

Active Founders

Yuhong Sun

Building the first open source Universal-Search/Chat-with-Docs across all workplace tools such as Confluence, Google Drive, etc. My expertise is in Machine Learning (with specialization in NLP) as well as backend systems. Previously working on Deep Learning for NL-to-SQL and Semantic Search at Alation, inc.

Yuhong Sun
Yuhong Sun

Chris Weaver

Former TL on the anti-fraud engine team at Robinhood. In a previous life, was a top-100 NA League of Legends player.

Chris Weaver
Chris Weaver

Company Launches

TL;DR: Imagine a secure ChatGPT with access to your team’s specific knowledge. As modern teams grow, knowledge becomes dispersed and hard to find. By giving Large Language Models access to parse your knowledge base, Danswer can instantly filter through millions of documents to only surface what is relevant.

Try for Free Here: https://app.danswer.ai/

Get Started On-Prem: https://docs.danswer.dev/quickstart

We’ve seen teams use Danswer to improve efficiencies from:

  • Improving turnaround times for support by reducing the time taken to find relevant documentation.
  • Helping sales teams get customer context instantly by combing through calls and notes.
  • Preventing lost engineering time to answering cross-team questions, building duplicate features due to the inability to surface old tickets or code merges, and helping on-call resolve critical issues faster by providing the complete history of an error in one place
  • Self-serving onboarding for new members who don’t know where to find information.

Our Differentiation:

Danswer is open source. You have complete transparency into what is going on with your team’s documents. The deployment of Danswer and your data are completely in your control (we do also offer a Cloud version if you prefer being hands-off). You can hook Danswer up with any LLM of your choice or one in-house if you prefer.

Danswer uses the latest methods in information retrieval to make the most of the revolution that is happening in the natural language processing space. Specifically, we use prefix-aware embedding models trained with contrastive loss + BM25 Hybrid Search with reranking options and additional pipeline improvements such as query expansion etc.

You can also use Danswer to directly get answers in Slack (MS Teams, Discord to come). Many teams have been using this to auto-answer cross-team questions.

Our Ask/Offer:

Reach out to us if you’re interested in using Danswer, you can schedule a call with us on Cal. Alternatively, you can join our Slack here — we’d love to hear from you!

We’re already deployed at some outstanding teams and we’re looking to accelerate. If you’re open to closing a deal with us quickly, we can offer you some significant discounts.