@luismiguelbarreiros Thanks for reporting! Because of the limit of the prompt size, it will never be able to know all the pages, but through prompt engineering it should be possible to make it explain itself in these cases better
@homer_meng Limit on characters but trying to split between paragraphs or sentences. But I should switch to token as a measure soon, there are a lot of Chinese users with a different token/character ratio.
this is soo cool, it can understand journal in Bahasa
gonna help a lot when trying to search information on journals
Anyway, I presume that, in either case, pdf chat always reads all the 200 pages of the document, even in the case when we ask him about the total amount of pages and he answers that it doesn’t know or keeps telling us that only knows pages 1, 3, and 13, for instance? Am I right?
Many thanks for your answer
Very nice, I discover if I upload a book in English for example and “chat with it” on Spanish, it answers me in Spanish! Amazing!
Hi, I can save my PDF to chat with it later?
I’m having this trouble too. How can I prompt it to actually read any of the document? It claims it can’t even read page 62, which it’s (randomly) told me it possesses. Thanks in advance. PS if it helps, this is a 121-page document.
I had a similar problem. Asked about the references and was answered that they were not available for analysis.
@xathis hello Mathis, I am a duke cs graduate with alternative medicine background. Working on LLM to against autism(ASD)， i found i need to have chatpdf like capability to index and semantically search my own 40MB txt files.However my chat bot (thru langchain)'s performance is good for English but poor for Japanaese and Chinese. Would you pls share some detail per your practice on loader、splitter and Embeddings ? Thank you so much！
ChatPDF is simply amazing! Your use of OpenAI GPT3 (right?) to construct this website is a true testament to their technical skill and creative problem-solving abilities. Uploading an entire PDF and being able to converse with it is nothing short of magical. Man! you deserves much more recognition for its ingenuity and innovation. Keep up the outstanding work!
Hello, is your background python or node
Can you connect GPT & Pdf so when there is a lack of data in the pdf i can ask the questions the gpt at same time, it’s like on your GPT mode,
This is really a great solution you have built! What I would be interested in is how you extract the text from the PDF documents?
I use OCR in another use case but the solution you built is much faster and performant! Big compliments for the great tool!
Hi! Thanks for bringing up this service. It is excellent for all Academic researchers (specially in social siences). It is helping me every single minute. Thank you
Need to paid? The free version too limited to use
One of my favorite prompts is “Make a questionary with answers to study”
Go and try.
I would love to know how you’re parsing the PDF file. Very nice work btw, love the idea.
how can i use free chatpdf without limited with api key?
It is a good app to assist reading, but to get the details and have your own thought, it is better read the PDF by yourself.