Subscribe
Sign in
Ragie
p/ragie-audio-videoBuild RAG applications on your user data
Start new thread
Ryan Gilbert

Ragie - Fully managed RAG-as-a-Service for developers

by
Ragie is a fully managed RAG-as-a-Service built for developers, offering easy-to-use APIs/SDKs, instant connectivity to Google Drive/Notion/and more, and advanced features like summary index and hybrid search to help your app deliver state-of-the art GenAI.

Add a comment

Replies

Best
Bob Remeika
Maker
📌
Hi Product Hunt! 😸 We're excited to officially launch Ragie on Product Hunt right on the heels of announcing our $5.5M seed round! So... what is Ragie? ✅ fully managed RAG-as-a-Service for developers 🤖 easy to use APIs and SDKs get you started in minutes 🔁 connectors that keep your application synchronized with data sources like Google Drive, Notion, Confluence and more Ragie enables developers to build AI applications connected to their own data in record time. At our core, we are a data ingest pipeline that extracts, chunks and indexes your data when you send us documents. You can send documents via our API or you can simply connect your application using one of our many connectors. When you are ready to use your chunks for generation, Ragie has a simple retrieval API. Our retrieval API uses re-ranking to make sure that you always get the best chunks and avoid hallucinations in your application. We're also working on hybrid search which should be released next week. We think one of the benefits of using Ragie is the fact that your application automatically improves as Ragie improves. With Ragie, you can focus on building your application instead of a RAG pipeline. I'll be in the comments all day and would love to hear what you think!
Phelipe Martin
@foobarfighter "Ragie offers simple, straightforward pricing without setup fees, hidden costs, or surprises." -> Book a demo You lost me here.
Mohammed Rafiq
Maker
Hi Product Hunt! We are super excited to launch Ragie on Product Hunt. Ragie can handle a wide variety of data types during ingest including PDFs and PowerPoints which go through a pipeline of extraction steps including OCR and LLM based ones so we can get the most information out of the data. Each data type - text, json, images, tables go through their own chunking strategy before they are embedded and indexed in a vector db. We build multiple semantic and keyword indexes out of the box which powers our powerful retrieval API which seamlessly picks the best combination of semantic, hierarchical and hybrid search indexes for the most accurate and relevant results. With Ragie, you can rest assured that our combination of extraction, chunking and indexing will always give you the best retrieval results out of the box without you having to worry about how to build and maintain a RAG pipeline. If you have any questions for us, please let us know.
Ryan Gilbert
Hunter
Congrats on the launch @foobarfighter and @mirafiq 🙌 Q: What's been the most challenging part of building Ragie up until this point?
Bob Remeika
Maker
@mirafiq @ryangilbert working on Ragie has been fast and furious. I think the hardest part up until now has been keeping up with the pace of the industry and keeping up with the demand to connect so many data sources. @mirafiq and I had both built several RAG applications prior to landing on the idea for Ragie. To be quite honest, it was really hard to build the quality of application that we wanted because we spent so much time building infrastructure and researching techniques for RAG instead of building our application. We also needed a lot of knowledge base data for our applications and every integration took a lot of time. Once we realized that the RAG infrastructure wasn't all that unique in each of our applications the idea for Ragie felt obvious. We raised a decent sized seed round so we could make sure that we had enough resources to really go after the RAG use case and build a great product for our customers.
Jeff Morris Jr.
@foobarfighter and team have built the best "RAG-as-a-service" product. One of the best engineering teams we have ever worked with. Maybe more surprising to hear... As a venture fund, we use Ragie internally at Chapter One for custom search engine that connects our Google Suite & other internal data to LLMs. So we are 2x believers... we invested and our developers actually use & love the product!
Bob Remeika
Maker
@jmj thanks so much! Really cool that you guys are using the product and we're able to add real value.
Ema Elisi
Congratulations on the launch of Ragie, @foobarfighter! 🎉 It's amazing to see how you've streamlined RAG-as-a-Service for developers. The features you’ve highlighted, like the easy-to-use APIs and instant connectivity to platforms like Google Drive and Notion, are game-changers for anyone looking to integrate AI into their applications quickly. Your emphasis on automated improvements in the application as Ragie evolves is a huge plus for developers focused on innovation without getting bogged down by the underlying complexities. The re-ranking feature for the retrieval API sounds promising; it could significantly enhance the user experience by minimizing hallucinations. I'm excited to see how Ragie will empower developers to create next-gen AI applications. Looking forward to seeing what the community thinks and what’s next on your roadmap. Best of luck with the growth post-launch!
Bob Remeika
Maker
@ema_elisi Thank you so much. Let us know if you think there's anything we can help you with!
Gary Der
Maker
@ema_elisi Thank you! Creating an great DX has been one of our north stars as we continue to build out and improve Ragie. We're always open to feedback so please reach out if you have any questions.
Bryan
Congrats on the launch @foobarfighter! 🎉 I’m curious about the retrieval API—how does it handle different data formats? Looking forward to the hybrid search feature too!
Bob Remeika
Maker
@dance17219 thanks for taking the time to check us out. On ingest, we usually convert media to text. So if you send us an image, we'll process that image using an LLM describe step and generate a text description. During retrieval, that image will be searchable via semantic search (and very soon keyword search). Upcoming, our plan is to return the original binary assets as base64 encoded chunks so they can be used in applications. If you have any feedback on how we're thinking about this, I would love to hear it.
Alexander Scott Williams
The way it handles data ingestion and retrieval is impressiv. Can’t wait to see how the hybrid search feature rolls out next week!
Bob Remeika
Maker
@alexanderscottwilliams thank you so much. Hybrid search will help applications avoid hallucinations but even without it the results are outstanding. Let us know if we can help you with RAG!
Robert Thomas
Hey there, Ragie looks super cool! Quick question - do I need to be familiar with data pipelines to get started? I’m a product manager and while I love AI, my coding skills are pretty basic. Would love to know if that might be an issue. thanks!
Bob Remeika
Maker
@robertthomas2 we intentionally built Ragie for people that did not want to deal with data pipelines. If you are technical enough to use AI tools to generate REST requests then you are probably technical enough to use Ragie's basic features. Ragie is built for developers so we support complex use cases but the simple cases are also supported. Give it a try :)
Matt Kauffman
Maker
@robertthomas2 we also have typescript (https://github.com/ragieai/ragie...) and python (https://github.com/ragieai/ragie...) SDKs if you're familiar with those languages. Adding a file to Ragie with a SDK is straightforward. In typescript it's just a couple lines. ``` const result = await ragie.documents.create({ file: await openAsBlob("./sample-file"), }); ```
Akhil E
Very interesting, RAG improves the performance of generative AI and this makes it very easy to integrate RAG with LLM models. Does Ragie handle data tables like let's say sales data or financial data which will be stored as csv files or in SQL databases?
Matt Kauffman
Maker
@akhil_e currently we support tabular data from CSVs (as well as typical spreadsheet formats). We'll definitely get to direct raw SQL support but it's not there yet. For now you can export SQL tables as CSVs and we can import the data that way.
Mckay Wrigley
Ragie is unbelievable! I’ve started to use it for all my retrieval needs in production. And it has an A+ developer experience. Really, really excited for more people to start building with it - it’s a gem.
Bob Remeika
Maker
@mckay_wrigley thanks! Love to hear that feedback from customers using us in production.
Patricia Harris
I love how Ragie simplifies the AI integration process, especially with the easy-to-use APIs! The re-ranking retrieval API is such a smart feature—avoiding hallucinations is huge! Gotta say, I'm a bit jealous of how streamlined this tool makes everything, it feels like magic.Can't wait to see what hybrid search brings next week.
Bob Remeika
Maker
@patriciaharris thanks for checking us out. Re-rank is such a great example of how Ragie makes your application better with no maintenance. We are constantly working on the latest and greatest in RAG and AI and you get all of that for free in your application.
Alex Klarfeld
We love using Ragie. I absolutely did not want to go through the hassle of setting up our own RAG-infra, plus dealing with all the complexities around chunking. Ragie takes care of all this for me. Question: We currently do our own scraping and ETL to load data into Ragie, do you all have plans to expand your integrations suite and include support for scrapers and other data sources?
Bob Remeika
Maker
@aklarfeld yes! Scraping is something that we have been looking at. We think synchronizing your application with your data is one of the largest barriers to entry when it comes to RAG and we plan on releasing at least 1 new connector weekly.
Alex Klarfeld
@foobarfighter That's awesome ! I'm excited to see all these new integrations!
Kyrylo Silin
Hey Bob, How does Ragie handle large amounts of data or high-volume requests? Do you have plans to add support for more data sources beyond the current connectors? Congrats on the launch!
Bob Remeika
Maker
@kyrylosilin thanks! Our pipeline is pretty robust because from "day one" we started building for customers that were already in production. We handle ingress up to 25MB per file which is large enough for most books and we have an API that also downloads files up to several GBs. We also account for bursty workloads which I think is somewhat unique to our offering. We are adding tons of data sources which includes Jira as our next officially supported integration. Is there a data source that you would like to see?
Rigel Cunningham
I love how this tool offers straightforward integration and powerful features. It’s definitely going to be useful for enhancing app capabilities. Keep up the excellent work! 👍🏻
Gary Der
Maker
@rigel_cunningham Thank you Rigel!
Ghost Kitty
Comment Deleted
Bob Remeika
Maker
@zulkarnaim thanks! Let us know what you think!
Ghost Kitty
Comment Deleted
Thomson Nguyen
Congrats @foobarfighter and @mirafiq on launching! Highly recommend any startup looking to kick the tires on setting up a RAG to check these folks out.
Adam Christian
This is incredibly cool @foobarfighter, congrats to you and your team - I look forward to watching this blow up. 🎈
Bob Remeika
Maker
@adamchristian thanks for the support!
Aaron Cort
Love Ragie!!!! Killer product 🤘💪🔥.
Bob Remeika
Maker
@aaron_cort1 thanks Aaron!
Dorothy Thompson
The integration with popular data sources is a huge plus.
Matt Kauffman
Maker
@dorothy_thompson our data connectors have been pretty popular! Let us know if there is a data source that you'd like to see us add.
drew dillon
Congrats @foobarfighter and Ragie team!!
Bob Remeika
Maker
Thanks @drewdil... appreciate the support!