Yes, you can use ChatGPT to summarize long documents for free. As of July 2026, ChatGPT’s free plan lets you upload up to 3 files per day (PDF, Word, TXT, and more), and each file can contain up to 2 million tokens of text — more than most books. For very long or accuracy-critical documents, splitting the text into chunks and summarizing section by section produces noticeably better results than a single “summarize this” request.
This guide covers both free methods step by step, the exact upload limits verified in July 2026, four copy-paste prompts with the output you should expect, and a simple rule for deciding which method fits your document.
Key Takeaways
- ChatGPT’s free plan allows 3 file uploads per day; each file is capped at 512 MB and 2 million tokens of text (as of July 2026).
- Uploads work for text-based PDFs and Word files — image-only scans are not supported, so OCR them first.
- For documents over roughly 30 pages, chunked summarization (summarize parts, then combine) beats one big prompt.
- Never send “summarize this” alone — specify audience, length, and format to get a usable summary on the first try.
In this guide:
- Can ChatGPT summarize long documents for free?
- Method 1: Upload the document (fastest)
- Method 2: Paste the document in chunks
- What are the best prompts to summarize a document?
- Which method should you use?
- FAQ
Can ChatGPT Summarize Long Documents for Free?
Yes. The free plan supports direct file uploads — up to 3 per day as of July 2026 — and each text document can hold up to 2 million tokens (roughly 1.5 million words). The practical constraints are the daily upload count and the free plan’s smaller context window, not file size. Here is how the limits compare:
| Limit (as of July 2026) | Free plan | Paid plans (Plus and up) |
|---|---|---|
| File uploads | 3 per day | Much higher (roughly 80 per 3 hours) |
| Max file size | 512 MB | 512 MB |
| Max text per file | 2M tokens | 2M tokens |
| Context window (how much it holds in one conversation) | ~16K tokens (~12,000 words) | 32K–256K tokens depending on model |
| Image-only scanned PDFs | Not supported | Not supported |
These numbers come from OpenAI’s official File Uploads FAQ; upload counts can be reduced during peak hours, so treat 3 per day as the baseline. One nuance matters for summaries: with very long files, ChatGPT does not always read every page in one pass — it searches the file for relevant sections. That is why the chunking method below exists.
Method 1: Upload the Document Directly (Fastest)
Uploading is the fastest free method: attach the file, send one structured prompt, and you have a summary in under a minute. It works for PDF, DOCX, TXT, and PPTX files with selectable text. Use it when you need a general overview and have one of your 3 daily uploads to spare.
- Go to chatgpt.com and sign in — a free account is enough.
- Click the + (plus) icon on the left side of the message box and choose Add photos & files.
- Select your document. Wait until the file card appears above the message box — sending before it finishes processing produces a reply based on a partial file.
- Paste a structured prompt (Prompt 1 below). Do not just type “summarize this” — you will get a vague paragraph and waste the upload.
- Ask follow-ups in the same conversation: “Expand point 3,” “List every deadline mentioned,” “Quote the exact wording on termination.” Follow-ups do not count as new uploads.
- Quick check: is the text selectable? Open the PDF and try to highlight a sentence. If you cannot, it is an image-only scan — upload it to Google Drive and open it with Google Docs first, which runs free OCR, then download the text.
Method 2: Paste the Document in Chunks (No Upload Needed)
Chunking — summarize each section separately, then combine the mini-summaries — costs more time but covers every page, works when you have used your 3 daily uploads, and stays inside the free plan’s ~16K-token context window. It is the method to use when missing a clause or a number is not acceptable.
- Split the document into sections of roughly 2,000–3,000 words each. Natural boundaries (chapters, headings) beat arbitrary cuts.
- Paste each section with the chunk prompt (Prompt 3 below). You get a ~150-word mini-summary per section.
- Start a new conversation, paste all mini-summaries in order, and send the combine prompt (Prompt 4). A fresh chat avoids running out of context.
- Spot-check one detail per section against the original — two minutes of checking catches most drift.
What Are the Best Prompts to Summarize a Document?
The best summarization prompts specify four things: audience, length, format, and what to preserve (numbers, names, conclusions). These four are copy-paste ready — replace the bracketed parts and use them as-is.
Prompt 1 — structured summary of an uploaded file:
Summarize the attached document for a busy reader. Format: 1. One-sentence overview 2. The 5 most important points (one line each) 3. Key numbers, dates, and names mentioned 4. 3 conclusions or action items Keep the total under 300 words. If any section of the document is unclear or missing, say so instead of guessing.
Expected output: a scannable, sectioned summary you can paste into notes or an email — not a wall of text. The last line noticeably reduces made-up details.
Prompt 2 — executive summary for someone else:
Write an executive summary of the attached document for [a manager who has not read it]. Length: 150 words. Tone: neutral and factual. End with one sentence on what decision this document supports or requires.
Expected output: a single tight paragraph plus a decision line — useful when the summary is going to your boss rather than your own notes.
Prompt 3 — summarize one chunk (Method 2):
This is part [X] of [N] of a longer document. Summarize THIS PART ONLY in up to 150 words. Preserve: key claims, numbers, names, and conclusions. Do not add information from outside this text. [paste section text here]
Expected output: a dense mini-summary per section. The “THIS PART ONLY” line stops ChatGPT from speculating about the rest of the document.
Prompt 4 — combine the mini-summaries (Method 2, final step):
Below are section summaries of one document, in order. Combine them into a single summary with: - a one-paragraph overview - 5 to 7 key points - a short conclusion Merge overlapping points instead of repeating them. [paste all mini-summaries here]
Expected output: one coherent summary of the entire document, built from parts ChatGPT actually read in full — the payoff of the chunking method.
Which Method Should You Use?
Upload when speed matters and an overview is enough; chunk when completeness matters or you are out of uploads. In practice the 30-page mark is a good switching point — below it, a single upload with Prompt 1 is reliable; above it, retrieval starts skipping content and chunking earns its extra effort.
- Choose Method 1 (upload) if the file has selectable text, is under ~30 pages, you have uploads left today, and a solid overview is the goal.
- Choose Method 2 (chunks) if the document is very long or dense, every section must be covered (contracts, research papers, reports you will be quizzed on), or you have hit the 3-uploads-per-day limit.
- Neither works directly on image-only scanned PDFs — run free OCR first (Google Drive → Open with Google Docs), then use either method on the extracted text.
FAQ: Summarizing Documents with ChatGPT
Can ChatGPT summarize a PDF for free?
Yes — as of July 2026 the free plan allows 3 file uploads per day, and text-based PDFs are fully supported. Image-only scanned PDFs are not: ChatGPT reads only the text layer, so OCR scans first (Google Docs does this for free), then upload or paste the result.
How long a document can ChatGPT summarize at once?
Technically, an uploaded file can contain up to 2 million tokens — well over a thousand pages. Practically, quality drops long before that because ChatGPT retrieves sections rather than reading everything. Past roughly 30 pages, use the chunking method if you need full coverage.
Does ChatGPT read the entire uploaded document?
Not always. For long files it searches for passages relevant to your prompt, which is fine for overviews but can miss details buried mid-document. Asking section-specific follow-ups (“Summarize chapter 4”) or chunking the text yourself forces complete coverage.
Summarizing documents is one of the highest-leverage AI habits you can build — it is the same principle behind our guide to 5 practical ways to automate repetitive work with AI. And once you have a summary, the most common next step is replying to whoever sent the document: these 10 ChatGPT prompts for professional emails handle that half of the job.