To chat with a PDF, open AI PDF, add your document, and type a question — the assistant answers from the text of that file, with the pages open beside the conversation so you can check what it tells you.
It also does the work, not only the reading. Ask it to compress, merge, delete pages or convert the document and it runs the same tools as the rest of the site, then hands the finished files back in the chat.
Here is what this covers:
- asking a question and getting an answer from your own document
- the questions worth asking a long file
- letting the assistant edit and convert the PDF for you
- chaining several steps into one request
- what to do when the answers are thin
- what happens to your document afterwards
Ask a Question and Read the Answer
The whole flow is three steps:
- Open AI PDF and add your document — drag it in, choose it from your computer, or pull it from Google Drive or Dropbox.
- Wait a moment while it uploads. The document opens on the left, the chat on the right.
- Type your question and press Enter. The answer arrives as it is written, so you can start reading before it finishes.
Shift + Enter gives you a new line instead of sending, and the send button turns into a stop button while an answer is streaming — useful when you realise halfway through that you asked the wrong thing.
Because the pages stay on screen next to the answer, you can scroll to whatever the assistant refers to and confirm it yourself. That matters more than any promise about accuracy: the document is right there.
What to Ask a Long Document
The assistant is at its best on the documents nobody wants to read end to end — a lease, a tender, an insurance policy, a 90-page user manual, a research paper.
Questions that work well:
- Summarize this document in bullet points — the fastest way to find out whether the file is the one you wanted.
- What are the key points? — the built-in suggestion, and a reasonable first move on anything unfamiliar.
- What are the payment terms? — specific clauses, dates, figures and names come back with the surrounding context.
- Where is X covered? — ask where a topic lives and then scroll the viewer to that part.
- What does section 4 actually mean? — plain-language restatements of dense passages.
Answers are drawn from the pages you uploaded. When something simply isn't in the document, the assistant says so rather than inventing it — which is the answer you want from a contract review.
Let the Assistant Do the Work
The conversation is not read-only. Ask for a change to the file and it calls the same processing tools that power the tool pages, shows what it is running, and lists the finished files with a download link on each.
| Ask for | What it runs |
|---|---|
| "Compress this" | Compress PDF |
| "Split this into separate files" | Split PDF |
| "Delete pages 3 to 5" | page deletion |
| "Pull out page 2" | page extraction |
| "Merge these" | Merge PDF |
| "Convert this to Word" | PDF to Word and the other Office formats |
| "Turn these pages into images" | PDF to JPG, and images back to PDF |
The files it produces are downloaded from the chat itself. When a request produces a long list of outputs, the chat shows the first ten and points you at the tool page for the rest.
Chain Several Steps in One Message
This is where a chat beats a row of buttons. You can describe the whole job in one sentence and let the assistant sequence it:
- "Delete page 2 and then compress"
- "Merge these PDFs and then compress"
- "Split out the signature page and convert it to Word"
Each step appears in the conversation as it runs, so you can see where a chain stopped if one part fails. Doing the same by hand means three tool pages and three round-trips of uploading and downloading.
Work With More Than One Document
Add several PDFs at the start, or use Add file in the chat header to bring another one into a conversation already underway. The header shows how many files are attached, and questions are answered across all of them — handy for comparing two versions of a contract, or for merging a set you are still reading.
Switch Document clears the workspace and starts over with something else.
When the Answers Are Thin
A few things account for most disappointing sessions:
The PDF is a scan. Answers come from the text stored in the file. A page that is really a photograph of a page has no text layer, so there is nothing to read — the assistant will have very little to work with. Documents exported from Word, Excel, or a web page are the best case.
The question is broader than the document. Ask about something the file doesn't cover and you will be told it isn't there. That is the assistant behaving correctly, not failing.
A step didn't run. If a tool errors, the chat says so and suggests trying again or using the tool page directly — the tool pages take the same file and are unaffected by whatever went wrong in the conversation.
You closed the tab. The conversation lives in the page, not in an account. Closing it ends the thread; the document can be picked up again, but the questions and answers are gone.
What Happens to Your Document
Your file travels over an encrypted connection and is permanently deleted from our servers within a few hours. The conversation itself never leaves your browser tab.
There is nothing to install — AI PDF runs in the browser you already have, on a laptop or a phone, and you can start without creating an account. If you work through documents regularly, a plan lifts the limits.
When you're done reading, the rest of the toolbox is one click away: compress the file so it fits in an email, split it into the parts people actually need, or convert it to Word to edit what you found. If size is your recurring headache, compressing PDFs is the fix for most of it.
Frequently Asked Questions
How do I chat with a PDF online?
Add one or more PDFs to AI PDF. Once they upload, the document opens next to a chat window — type your question and the assistant answers using the text of those files.
What can I ask about my document?
Ask for a summary, the key points of a section, specific dates and figures, or where a topic is covered. The assistant answers from the document and says so when something is not in it.
Can the assistant edit or convert the PDF for me?
Yes. Ask it to compress, merge, delete pages, convert, or chain several of those steps in one message. It runs the same tools the rest of the site uses and hands the finished files back in the chat to download.
Does it work with scanned PDFs?
Answers come from the text stored in the PDF, so a scanned page with no text layer gives the assistant little to work with. Documents exported from Word, Excel or a web page work best.
Can I chat with several PDFs at once?
Yes. Add more than one file at the start, or use Add file during a conversation — questions are answered across everything attached, and you can ask for them to be merged from the same chat.
Is my document kept private?
Your file travels over an encrypted connection and is permanently deleted from our servers within a few hours. The conversation itself lives in your browser and is gone when you close the page.
