Chat with a Document

Talk to your scans, images and documents.

No languages match your search.

🗨️

Drag and drop, or paste a screenshot

Image, PDF, Word (DOCX) or text - drop, browse, paste, or use a URL

🔒 Your files are processed privately and deleted automatically.

Chat with a Document lets you talk to anything you can scan or photograph, a printed page, a screenshot, a receipt, a letter, a form, or a multi-page PDF. ocr.chat extracts the text first, then lets you search and chat with it: ask what it says, pull out a specific detail, or get a quick summary. It is conversational OCR, the recognition and the question-answering happen in one place, so you go from an image to answers without copy-pasting text into a separate chatbot.

It is for the moments when you have a document in front of you and a question about it, not time to read the whole thing. Snap a photo of a notice and ask what the deadline is. Upload a scanned manual and ask how to do one specific step. Drop in a report and ask for the three key points. Because answers are grounded in the text that was actually extracted, and each one cites the page it came from, you get something you can trust and verify rather than a generic guess.

There is nothing to install and no signup to try the OCR. Upload an image or PDF, watch the text appear beside your original, and start asking. Recognition covers 100+ languages, your files are deleted automatically after processing, and nothing is ever sold or shared. Document chat runs on a free account, with paid plans from $5/mo for more pages, longer documents, and unlimited questions.

How to chat with a document

1
Upload an image or PDF
Drag in a photo, paste a screenshot, or select a scan or PDF of the document you want to talk to.
2
Watch the text extract
ocr.chat reads the document and shows the recognized text next to your original so you can see what was captured.
3
Chat with it
Ask questions, request a summary, or search for a detail right in the chat panel beside the document.
4
Verify with citations
Answers cite the page they came from, so you can check anything against the source before you rely on it.

Common uses

  • Reading a photo of a notice, sign, or letter and asking what it actually requires of you.
  • Pulling a single fact, a date, an amount, a reference number, out of a scanned document fast.
  • Summarizing a multi-page report or handout into a few clear points before a meeting.
  • Understanding forms, contracts, and official paperwork without reading every line.
  • Studying from scanned notes and slides by asking questions instead of rereading them.
  • Digitizing and querying receipts, statements, and records for quick lookups.

Frequently asked questions

Images, screenshots, photos, scans, and PDFs all work. The text is extracted first, then you can ask questions about any of it.

A regular chatbot does not know your document. Here the answers come only from the text extracted from your file, with a page citation, so they are grounded in what you actually uploaded.

Printed scans work on the free engine; messy handwriting is best read with the Premium AI engine. Either way the recognized text is what you then chat with.

The assistant answers only from your document and cites the page it used, and your original is shown beside the text so you can verify. If something is not in the document, it says so.

Recognition covers 100+ languages, and you can ask and receive answers in your own language.

Yes. The conversation is threaded, so follow-ups build on earlier answers about the same document.

Extracting text is free with no signup. Chatting with the document uses a free account, which also keeps your history and gives you more pages monthly.

Short and medium documents work on the free tier; larger documents and more questions are included in paid plans. Pages count toward your quota when read.

Yes. You can copy or export the full extracted text as plain text, Markdown, Word, or a searchable PDF in addition to chatting.

Your file is processed only to extract text and answer your questions, then deleted automatically. We never sell or share your documents.

Yes. Each answer references the page it drew from, so multi-page documents stay easy to check against the original.

Yes. Paid plans include a REST API for extracting text and automating document question-answering in your own apps.

Use this via the API

Run this tool programmatically with a single POST. Authenticate with the API token from your account page.

curl -X POST https://ocr.chat/api/v1/ocr/ \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -F "file=@your-file.png" \
  -F "tool=document"

Files of 5 pages or fewer return the result inline; otherwise poll the job, then download it as md:

curl -L "https://ocr.chat/api/v1/ocr/JOB_UUID/download/?format=md" \
  -H "Authorization: Bearer YOUR_API_TOKEN" -o result.md

Then ask questions about the document, with answers cited to the page:

curl -X POST https://ocr.chat/api/v1/chat/JOB_UUID/ \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"message": "Summarize this document"}'
Read the API docs →

More OCR tools

🖼️ Image to Text 📄 PDF to Text 📝 PDF to Word 🗎 Image to Word ⬇️ PDF to Markdown 📊 Table Extractor 🔍 Searchable PDF ✍️ Handwriting to Text ➗ Math to LaTeX 🧾 Receipt Scanner 💳 Invoice OCR 🌍 Translate a Document 🔤 Word to Text 📃 Word to PDF 📈 PDF to Excel 🧮 Image to Excel 🌐 Image Translator 👁️ Describe Image ♿ Alt Text Generator 💬 Chat with PDF
Rate this page
★ ★ ★ ★ ★
5.0/5 (0)

What could we improve? Your feedback helps us fix issues.