Quivr (YC W24)

Quivr (YC W24) just open-sourced a way to efficiently parse PDFs for optimal LLM Ingestion! Making RAG applications better! 🔥 We used unstructured.io under the hood, powered by some algorithms and LLMs to clean the files and improve it. For example, we use OpenAI GPT4o for improving the tables that are extracted - you could use any model, even local ones. Link for the open-source project in the comments 👇

Quivr (YC W24) is recruiting not 1, not 2 but 3 Founding Engineers! 🔥 We are coming back from Y Combinator next week and are looking for our founding team! 👇 TL;DR: Quivr is an open-source RAG framework with 31k+ Github Stars, backed by Y Combinator. Here is a sneak peek of what we are looking for: 1️⃣ Head of AI We’ve got some big algorithms to cook up for our clients and the community! We need to make Quivr a standard in the RAG-powered agentic framework space. It will involve some IPO, ETL, retrieval, embedding, maths & logics to cook up some good algorithms. 2️⃣ Backend Engineer Quivr’s API and backend architecture need to be modular, robust and fast. We need to make this into a framework that everyone can use and run in a few lines of codes. Think of Quivr as the LangChain for RAG-powered agents in production with a sleek UI. 3️⃣ AI Engineer RAG and agents! We’ve got some tricky rag-powered agents to implement using Quivr. Combine Ingestion, Embedding, Vectors, Fine-tuning to implement powerful agents. You'll be working closely with Antoine Dewez, Brian Laleye & I! ❗ We are hiring in Paris, France ❗ English speakers welcomed! Remote-friendly 🏠 And of course, you need to be an open-source enthusiast! We have a fully open-source stack Supabase, PostHog, LangChain, FastAPI... if you are interested, drop us an email: founders(at)quivr(dot)app And if you are not, tag a friend or like this post so your friends can see it!

I'm very proud to announce our latest feature: Chat with your Database 🔥 Allowing anyone to easily chat with a database! Here is how you can use it: TL;DR: Quivr (YC W24) we allow anyone to easily connect a database so that you can chat with it. 1️⃣ Ask one of your devs to create a brain and connect it to your database 2️⃣ He can then share it to you 3️⃣ Ask any questions you want on your database This allows you to easily ask any questions on database, this used to be very complicated and needed to create an API, a frontend and any changes would take days. Now you can do it in less than 30 seconds (proof in video 🤣) And the best about it ?! It is open-source ! Link in the comment.

I will develop for 10 companies a production ready GenAI assistant for your business case for FREE using Quivr (YC W24) 🔥 💵 What do you need ? Chat with PDF, Excels, Database or agents 👇 TL;DR: Quivr (YC W24) is an open-source RAG framework that makes it easy to create and deploy "Second brains" that solve issues you encounter in your company. 1️⃣ I will develop a custom brain on Quivr that will perform your use case 2️⃣ I will deploy it in production either on Quivr.app or on your custom instance 3️⃣ You can own the code and no data goes outside of your infrastructure thanks to LangChain, Supabase, Mistral AI. We can deploy Quivr to any cloud provider and no data leaves your datacenter. If you are interested, please comment and connect with me so I can contact you ! 🔥 Limited to 10 companies !

Last year was the time of POC, this year it is the time to go to Production with GenAI. You can now create & deploy custom agents in minutes with Quivr (YC W24) 🔥 Here is how: TL;DR: You can create custom agents specific to your business use case in minutes with Quivr and deploy it to production in a robust framework. 1️⃣ Download Quivr from the open-source repository (link in the comments) 2️⃣ Create your own agent using LangChain 3️⃣ Deploy it to production with Quivr (YC W24) Here is a small video that shows how you can do it 👇 We implemented custom brains such as Notion, Excel, SQL that allows your users to talk to their data wherever they are. Have fun! The code is open-source so don't hesitate to improve it 😘

I would have never thought that 9 months ago when we started working on Quivr (YC W24) that we would be recognised as one of the best projects from Andreessen Horowitz in AI 🔥

Quivr (YC W24) is building an AI, open-source platform (with 28k+ GitHub stars!) that enables businesses to connect tools, documents, APIs, and databases for direct chat and content generation. Employees spend up to 20% of their time searching for company information, encountering challenges like repetitive requests, difficulty in finding necessary details, and uncertainty about the existence of needed information within the company. These inefficiencies not only delay tasks but also lead to decisions made without access to pertinent information, hindering effective communication and collaboration across teams. Founded by Stan Girard, Antoine Dewez, and Brian Laleye, Quivr enables enterprises to integrate their company data, facilitating easy retrieval of information and generation of new content. It streamlines tasks such as writing emails without the need for contextual input, extracting actionable data from large databases, or summarizing extensive documents to capture their essence. Congrats Stan, Antoine, and Brian on the launch!

Do you want superpowers with your Excels? 🪄 We are working on something amazing to let anyone talk with their Excel files 👇 TL;DR: Say interested in the comments if you want access to the Beta 🥷 How does it work? 👀 1️⃣ You ingest your CSV into Quivr (YC W24) 2️⃣ You ask the question 3️⃣ Abracadabra 🪄 Remember all these times you spent trying to write Excel formulas?! No more! This feature is still in preview and we are looking for beta testers. If you'd like to try it, simply leave a comment down below 👇

I can chat with all the essays of Paul Graham (co-founder of Y Combinator) to ask him help on Quivr (YC W24)! Here is how you can do it too 👇 Simply go to https://quivr.app and create an account (5 secs with Google) and ask the question: What are Paul Graham advices to achieve PMF ? Quivr will then find Paul Graham brain that we created and ask the question to it. If it doesn’t answer with Paul Grahams brain simply do a "@ Paul Graham" Hope it helps you find new ideas ! Enjoy ! And have fun finding PMF while talking to your PG companion!

It took me less than a minute to analyse & summarise a 151 pages long public RFI/RFP with Quivr (YC W24) ! 🔥 Here is the 60 seconds process: TL;DR: Quivr is open-source & can be installed locally. The hosted version is limited to 20/30 pages long documents in the free version. 1️⃣ Go to https://quivr.app 2️⃣ Create a brain 3️⃣ Add your RFI/RFP 4️⃣ Ask questions to it 5️⃣ Enjoy the time you gained. It is that easy ! And you can even run it locally in your enterprise network to ensure maximum compliance. If this is a pain for your enterprise or yourself and like to use this feature, we'd love to catch a call with you: https://lnkd.in/d7baEc2R