More excellent context Dmitry - thank you. You are helping to reshape the way I go about research, and introducing exciting new tools that are available in the AI era.
Chatpdf is excellent for maintaining context and providing accurate information from the PDF. However, the only problem I experienced while using it, was limitation in offering only a few interactions. It kept popping an error of ..."Error limit..."
I recently stumbled upon ChatDocs and i feel it's the first of these apps that actually stands out way above the others and I'm interested to see if you think the same thing! Not only is the interface and UX way better than the others but i feel like the query algorithm is different and better than the rest of these LangChain based RAG apps! I feel like it may ingest the documents different and can answer more complicated and complex questions ... What do you think? Am i just hallucinating or is it actually different? I hate can only trial with 2x20 page files before they ask for money tho cause the documents i usually test with are much larger lol
To me, the favorite feature of ChatDocs is the ability to not stick to the context. Because I can read my PDF documents, but it's much more interesting to use the content in the PDF to explore the boundaries beyond a specific context. Have you tried InfraNodus already? What do you think?
I want to thank you for the help I have gone through all your videos. But I want to ask if you can take a pdf and do all of this steps from the upload step as I have problem uploading the document and getting the same result
Thank you! What is the problem you're having exactly? I'll be happy to help you! Also, if you want, our customer support specialist Thomas can jump on a video call with you and show you some tricks.
@@noduslabs I uploaded a document to visulize it gave me alot of nonsense like ho (for house), * (for some bullet point it thought important), pre (prerelease).. And so on
Yes! I need this, badly, kindly help me achieve this with a vault totaling 2 million+ words please, consider it a benchmark test of your capabilities. How the hell can I arrive at that desired context output?
@@noduslabs theme placement...there are specific themes for a presentation it represents, scattered throughout, and I want to see which points would be best to hit and cross reference them.....they are markdown files.
Good day, i meant to confirm for all this tools are they limited to chatting to one document only what about to more than one but in a folder, and also disclosing the source doc and page numbers?
You can chat with multiple documents and if you use the new extension on top, you can actually get the source data from which the content was generated, yes.
I need a program that can read one or several documents to answer questions in another uploaded document. I don't want to have to enter the questions manually. Any suggestions?
@@noduslabs I need to prepare materials to assist in the checking of manual work completed by humans. To do this wok manually would be more time intensive.
@@EdwardSantoro Sure but not to the extent where you're asking me how to basically automate everything that you do, including the process of automation itself. It's like I'd say: how can I automate education, so that instead of having to find the answers to questions myself, I'd have an AI that would do that for me. That's how your question sounded to me :)
Sure, well, first of all if it's a PDF then probably there's not 23 Mb of text, so you can simply use another tool to extract the text and upload it as a TXT file to bypass the limitation. Another thing is if it's actually 23 Mb of text. In this text, visual analysis of this doc is not going to be so interesting for you because you will see a graph that is too dense and there won't be meaningful clusters emerging. In this case, I'd suggest to work with smaller documents and discourses that are somehow manageable. We are working on a solution for bigger docs, but it probably won't have the visual aspect, just the main topics, clusters, with the ability to choose where you want to zoom in.
Yes, it's similar but specialized PDF processing tools based on AI are just better designed for working with longer documents and multiple ones as well.
@@noduslabs i understand, thank you for the answer. i have been comparing, chatgpt is the only one i pay for, so i cannot compaare with the paid versions of the others, and i found the answers of gpt better... but i dont knot, sth i still need to gather more comparables exemples
@@marcelo_.hs-procureyou're right, they might have improved it. at the beginning, ChatGPT had a lower context window but then they increased it so prob it's even better than ChatGPT. Why are you still paying for it btw if ChatGPT is better?
Yes, I understand your disappointment, but we're a small company and there are maybe 2 people using the tool in Polish, so it's simply not possible to us to program for it. However, you could auto-translate this doc to English in just a few seconds with DeepL (a great tool and also converts PDFs very quicklY) and then upload it to InfraNodus for analysis. Plus GPT-4 works better with English anyway. Of course, we will strive to accommodate our Polish users as well as soon as we are able to!
More excellent context Dmitry - thank you. You are helping to reshape the way I go about research, and introducing exciting new tools that are available in the AI era.
Chatpdf is excellent for maintaining context and providing accurate information from the PDF. However, the only problem I experienced while using it, was limitation in offering only a few interactions. It kept popping an error of ..."Error limit..."
I recently stumbled upon ChatDocs and i feel it's the first of these apps that actually stands out way above the others and I'm interested to see if you think the same thing! Not only is the interface and UX way better than the others but i feel like the query algorithm is different and better than the rest of these LangChain based RAG apps! I feel like it may ingest the documents different and can answer more complicated and complex questions ... What do you think? Am i just hallucinating or is it actually different? I hate can only trial with 2x20 page files before they ask for money tho cause the documents i usually test with are much larger lol
To me, the favorite feature of ChatDocs is the ability to not stick to the context. Because I can read my PDF documents, but it's much more interesting to use the content in the PDF to explore the boundaries beyond a specific context. Have you tried InfraNodus already? What do you think?
I want to thank you for the help I have gone through all your videos. But I want to ask if you can take a pdf and do all of this steps from the upload step as I have problem uploading the document and getting the same result
Thank you! What is the problem you're having exactly? I'll be happy to help you! Also, if you want, our customer support specialist Thomas can jump on a video call with you and show you some tricks.
@@noduslabs I uploaded a document to visulize it gave me alot of nonsense like ho (for house), * (for some bullet point it thought important), pre (prerelease).. And so on
Yes! I need this, badly, kindly help me achieve this with a vault totaling 2 million+ words please, consider it a benchmark test of your capabilities. How the hell can I arrive at that desired context output?
What kind of insights you want to get from your 2 Mln words and what kind of data you have?
@@noduslabs theme placement...there are specific themes for a presentation it represents, scattered throughout, and I want to see which points would be best to hit and cross reference them.....they are markdown files.
@@noduslabs I responded
@@noduslabs this one 🎯
Good day, i meant to confirm for all this tools are they limited to chatting to one document only what about to more than one but in a folder, and also disclosing the source doc and page numbers?
You can chat with multiple documents and if you use the new extension on top, you can actually get the source data from which the content was generated, yes.
I need a program that can read one or several documents to answer questions in another uploaded document. I don't want to have to enter the questions manually. Any suggestions?
Hahaha :) Why would we then need humans if we leave all of it to the machiens?
@@noduslabs I need to prepare materials to assist in the checking of manual work completed by humans. To do this wok manually would be more time intensive.
@@noduslabs Then what is the purpose of using AI at all, if not to automate lower-level processes?
@@EdwardSantoro Sure but not to the extent where you're asking me how to basically automate everything that you do, including the process of automation itself.
It's like I'd say: how can I automate education, so that instead of having to find the answers to questions myself, I'd have an AI that would do that for me. That's how your question sounded to me :)
hello it is great tool , but my concern is the limit size of document for example if i have document with 23MB what can i do
Sure, well, first of all if it's a PDF then probably there's not 23 Mb of text, so you can simply use another tool to extract the text and upload it as a TXT file to bypass the limitation.
Another thing is if it's actually 23 Mb of text. In this text, visual analysis of this doc is not going to be so interesting for you because you will see a graph that is too dense and there won't be meaningful clusters emerging. In this case, I'd suggest to work with smaller documents and discourses that are somehow manageable.
We are working on a solution for bigger docs, but it probably won't have the visual aspect, just the main topics, clusters, with the ability to choose where you want to zoom in.
why dont upload pdfs on chat gpt instead? is it the same as chat pdf?
Yes, it's similar but specialized PDF processing tools based on AI are just better designed for working with longer documents and multiple ones as well.
@@noduslabs i understand, thank you for the answer. i have been comparing, chatgpt is the only one i pay for, so i cannot compaare with the paid versions of the others, and i found the answers of gpt better... but i dont knot, sth i still need to gather more comparables exemples
@@marcelo_.hs-procureyou're right, they might have improved it. at the beginning, ChatGPT had a lower context window but then they increased it so prob it's even better than ChatGPT. Why are you still paying for it btw if ChatGPT is better?
I'm disappointed that Infranodus doesn't handle PDF in Polish well. Without it, the program becomes useless to me...
Yes, I understand your disappointment, but we're a small company and there are maybe 2 people using the tool in Polish, so it's simply not possible to us to program for it. However, you could auto-translate this doc to English in just a few seconds with DeepL (a great tool and also converts PDFs very quicklY) and then upload it to InfraNodus for analysis. Plus GPT-4 works better with English anyway.
Of course, we will strive to accommodate our Polish users as well as soon as we are able to!