AnyParser

Hey everyone 🎉 This is Rachel, Cofounder of Cambio. Every day, AI extract data from billions of documents or images. However, traditional OCR models often struggle with understanding semantic information, and they frequently miss crucial details in diverse document formats. That's why we've developed AnyParser—an LLM just for document extraction. With AnyParser, your AI applications can access a richer trove of information! What do our customers love about AnyParser? - ✅ **Json-mode**: Extract json-like key values from your document with AnyParser. - 🔐 **Multi-language**: We support common used languages beyond English, such as Arabic. - 📊 **Low latency**: AnyParser is powered by special LLM which yield 200 tokens per second on a single GPU, 5x faster than GPT4o. - 🛡**Privacy Protection**: Activate the "Remove Private Information" feature, and AnyParser will automatically redact P.I.I. (Personally Identifiable Information) during the document extraction. - 📈 **High Accuracy**: Bid farewell to jumbled tables and chaotic layouts that plague traditional OCR-based models. Ready to get started? - If you’re a developer working on RAG or LLM applications, get a free API key today! - If you want to test the model performance directly, try your hand directly in our AnyParser Sandbox! We’re here to answer your questions and discuss how we can help enhance your AI applications. Cheers, Team Cambio

Rachel this looks highly impressive can't wait to give this a go & unreal that you've managed to achieve such high speeds :) The security aspect of this is vital too, I always get concerned about putting in important documents to any platform so I'm glad you can give me that guarantee. Best of luck w/ the launch!

Congratulations on the launch! I'm really impressed with your product. 👏

An accurately working parsing tool is incredibly useful, and Cambio seems to deliver just that! I believe you could improve the conversion rate on your landing page by tweaking the initial design. Instead of showing the parsing interface upfront, consider displaying a drag-and-drop block with an upload button. This will make the page clearer, help to focus attention, and remove the confusion of empty state inteface. It’s also fascinating how a registration could be delayed during an onboarding flow. For instance, you could prompt users to register after they’ve uploaded a file, when they try to copy text, or after parsing three documents. This approach will reduce friction and let users see the tool’s value before committing.

Congrats! This product is really impressive. My favorite part is the definition of privacy protection, which is crucial for entering serious enterprise markets.

Sorry, but I tried cambioML multiple times, once with a The New York Times article and then a TechCrunch article and it failed both times. I have submitted feedback. Clearly, the engine needs more testing. 🤔 The good news is image parsing worked for a PNG chart. 😊

@rachel_hu Congrats launch, good work for LLM.