Question 1

Does 'document Q&A' mean any file type — Word, scanned images, plain text?

Accepted Answer

Not yet — and we want to be exact about it. Today this feature works on text-based PDF files specifically. It's called document Q&A because that's the search term people actually use and it describes the job (ask a document questions), but the runtime support right now is PDF only. Word documents and plain text files aren't accepted through this flow, and scanned or image-only PDFs return a 'no extractable text' message.

Question 2

How do I ask a document questions inside Telegram?

Accepted Answer

Open @vustSummaryBot, send your PDF as a document upload, and read the summary it returns. If the summary came from a PDF, a "💬 Ask about this PDF" button appears underneath. Tap it once, then send your question as a plain message — the bot treats it as a question about that specific document and answers from its text.

Question 3

How many questions can I ask, and does it cost anything?

Accepted Answer

Opening the chat and getting the first summary is free. Each follow-up question costs 1✦, up to 3 questions per document, and only within 30 minutes of the summary being delivered — the window doesn't extend when you ask a question. After 3 questions or 30 minutes, resend the document to start a fresh round.

Question 4

Does it answer from the document itself, or does it guess?

Accepted Answer

From the document. Every answer is grounded in the actual text you sent — not the open web, not general knowledge about what a typical document of that type usually says. If the question isn't answered anywhere in the document, the bot says it can't find it instead of inventing a plausible-sounding answer.

Question 5

Can it read a scanned document or a photo of a page?

Accepted Answer

No. The bot reads the selectable text layer of a PDF; it doesn't run OCR. A scanned document or an image-only PDF has no text to extract, so you'll get a message saying there's no extractable text, with a suggestion to paste the content directly as a chat message instead.

Question 6

What if my document isn't a PDF — will you expand to other formats?

Accepted Answer

Not something we can promise on a timeline, so we won't claim it here. What's true today: PDF documents, text-based, up to 3 questions each. If you have a Word or text file, converting it to PDF first is the current workaround — the resulting PDF works the same as any other in this flow.

Question 7

What kinds of documents work well for this — any real-world examples?

Accepted Answer

Anything text-heavy where you'd otherwise scan for one detail: rental or vendor agreements, insurance policies, research papers, syllabi, terms of service, meeting handouts exported as PDF. The common thread is a text-based PDF with the answer somewhere in its own words — the bot won't reason beyond what's on the page, so it fits fact-lookup questions better than open-ended interpretation.

Question 8

Can I ask a follow-up that builds on my previous question?

Accepted Answer

Within the same 30-minute window and 3-question cap, yes — each question is answered against the full document text, not just your prior question, so you can ask "what's the notice period" and then "does that apply to early termination too" and get a grounded answer to the second question as well, not a guess based on the first answer alone.

Ask a Document Questions — Grounded in Its Text.

'Document' means PDF today — we're saying so plainly

See the difference

The full flow, start to finish

A grounded answer, in-chat

Who asks a document questions in Telegram

From document to grounded answers

Send a document, ask it questions

What's live — and what's honest about it

PDF today, not every document type

No OCR

Bounded by design

Frequently asked questions

More document Q&A tools.

Ask your document. Get an answer from its text.