Researched
This analysis is based on documentation, public user reports, and vendor materials — not yet on our own hands-on testing. How we rate
Drag a file in, start asking
The whole product is in that sentence. Upload a document, type a question, get an answer with the page it came from.
No setup, no notebook to create, no configuration. Three seconds from file to first question. That frictionlessness is the entire reason ChatPDF has the user base it does, and it's genuinely hard to beat — competitors with better capabilities routinely lose on how many steps stand between you and an answer.
The page citations are the other thing worth crediting. Every response points at where it came from, which means checking the tool's work takes seconds instead of requiring you to re-read the source. Given how often AI tools produce confident wrong answers, that's not a small feature.
The pricing is genuinely unclear
Most tool pages state a price. This one has to state a range, because the sources don't agree and the disagreement looks real rather than sloppy.
Credible reviews published within a few months of each other report ChatPDF's paid tier at $5, $9.99, $14.99 and $19.99 a month. That's a four-fold spread.
What seems to be happening, based on the most detailed recent breakdowns:
| Tier | Reported price | Limits |
|---|
| Free | $0 | 2–3 PDFs/day, ~50 questions each, ~120 page cap |
| Plus | ~$5/mo | Substantially higher daily limits, ~500 page cap |
| Max | ~$15/mo | ~500 PDFs/day, ~2,000 page cap |
| iOS in-app | ~$14.99/mo, ~$89.99/yr | Apple App Store listing |
Two explanations fit the evidence. First, the tier structure changed during 2026 — older reviews describing a single $19.99 Plus plan predate a restructure into Plus and Max. Second, and this is well documented, the Apple in-app price differs from web checkout, which is common when a vendor prices in the App Store commission.
The practical advice: check the web checkout page and the app store separately, and buy on whichever is cheaper. That's an unsatisfying answer and it's the accurate one.
What the free tier actually covers
Two to three documents a day, around fifty questions each, and a page ceiling near 120.
For a student working through a handful of papers a week, that's the whole job. For someone checking a lease or a contract occasionally, likewise. The free tier here is a real product rather than a nag screen, which is worth saying because most tools in this space aren't.
Where it stops: literature screening at any volume, and long documents. A 300-page report won't upload. And a frequently reported irritation is that failed uploads sometimes still count against your daily allowance, so a botched attempt costs you one of your two or three.
The 2026 expansion
ChatPDF stopped being only about PDFs this year.
Format support now covers PowerPoint, ePub, Word, Markdown and plain text. Page limits on the top tier rose to around 2,000. Multi-file folders arrived, finally addressing the complaint that you could only interrogate one document at a time.
Then there's the rest: YouTube video chat, AI research features, flashcard generation, slide creation, an AI writer, an AI detector, mobile apps and a desktop app.
Whether this is good depends on your view of feature expansion. The original product was excellent because it did one thing with no friction. Adding a flashcard generator and an AI detector doesn't make the document chat worse, but it does suggest a company looking for growth in adjacent directions rather than depth in the original one. Meanwhile the most-requested fix — OCR — still isn't there.
The OCR problem
No OCR. This is the limitation that determines whether ChatPDF works for you.
A scanned page is an image. ChatPDF reads text. If your PDF came from a scanner, a phone camera or an archive digitization project, the tool sees nothing usable and returns nonsense or blanks.
For a lot of use cases this never comes up — modern documents are born digital. For anyone working with older material, court filings, medical records or anything photographed, it's disqualifying. Run files through an OCR step first, or pick a tool that handles it.
Accuracy on documents it can read is decent and degrades with length and structural complexity. Users report inconsistencies on dense material, and hallucinated citations appear in enough accounts to warrant caution. The citations mitigate this rather than eliminating it — you still need to check the cited page rather than trusting the summary.
How we researched this
We haven't run ChatPDF through a sustained research workload, and this page is desk research.
It's based on ChatPDF's own documentation and API limits, cross-referenced against independent reviews published between January and June 2026, plus Apple App Store listing metadata for the mobile pricing.
The research problem here was pricing, and it's worth being explicit about how we handled it. Rather than picking whichever figure appeared most often, we've presented the spread and offered the most plausible explanation for it. Several sources state a single price confidently; they contradict each other; at least one of them is describing a plan that has since changed. Anyone telling you ChatPDF costs exactly one amount either checked recently on one platform or is repeating a number they didn't verify.
What we can't tell you: how ChatPDF's accuracy compares against alternatives on your specific documents. Accuracy claims in this category deserve testing rather than aggregation, and we haven't tested.
Where it falls down
Scanned documents. Covered above. Hard stop.
Cross-document research. Multi-file folders help, but genuine synthesis across a body of sources isn't what this is built for. NotebookLM does that properly and costs nothing.
High-stakes accuracy. Contracts, medical documents, legal filings — use it to locate relevant sections, then read them. Don't rely on the summary.
Very large files. The 32MB ceiling bites before the page limit on anything image-heavy.
Academic literature at scale. Purpose-built research tools handle screening, citation and synthesis far better.
Against the alternatives
Against NotebookLM: better on capability and free. Multi-document synthesis, source grounding, and a genuinely strong research workflow. ChatPDF's only advantage is speed of getting started, and for regular research work that isn't enough.
Against Elicit: built for academic literature review specifically — screening papers, extracting data across studies, systematic review workflows. Different and much deeper job.
Against Consensus: answers research questions across published literature rather than a document you supply. Complementary rather than competing.
Against Claude or ChatGPT: both accept document uploads and reason over them better than ChatPDF does, with far larger context windows. If you already pay for one, you probably don't need this. ChatPDF's counterargument is the citations and the purpose-built interface.
Against Hebbia: enterprise document analysis for finance and legal, at enterprise prices. Vastly more capable and not aimed at individuals.
Against ScholarAI: research-paper focused with academic database access.
Pricing 2026
| Plan | Reported cost | Limits |
|---|
| Free | $0 | 2–3 documents/day, ~50 questions each, ~120 pages |
| Plus | ~$5/mo | Higher daily limits, ~500 page cap |
| Max | ~$15/mo | ~500 documents/day, ~2,000 page cap, 32MB per file |
| iOS in-app | ~$14.99/mo · ~$89.99/yr | App Store pricing differs from web |
Checked August 2026. ChatPDF's paid pricing is reported inconsistently across credible sources at $5, $9.99, $14.99 and $19.99, likely reflecting both a tier restructure during 2026 and a difference between web checkout and Apple in-app pricing. Format support expanded in 2026 to include Word, PowerPoint, ePub and Markdown. OCR is not supported. Verify pricing on both the web checkout and your app store before subscribing.
Check web and app store prices separately. They differ, and the gap is large enough to matter.
Exhaust the free tier before paying. Two to three documents a day covers more use than people expect, and hitting the limit repeatedly is the signal to upgrade.
OCR your scans first. ChatPDF cannot read them and never tells you clearly that this is why the answers are wrong.
Always open the cited page. The citations are the product's best feature and they exist to be used, not trusted blindly.
Our Verdict
ChatPDF is the fastest way to interrogate a single document, and speed of getting started is a genuinely underrated feature. Drag, ask, get an answer with a page number. No competitor with better capabilities matches that lack of friction, and for a lot of people that trade is correct.
The free tier is honest. Two to three documents a day with real question limits is enough for light use rather than being a teaser. That's rarer than it should be.
The pricing is a mess to pin down. A four-fold spread across credible sources, a tier restructure, and different prices on web versus mobile. None of this is deceptive exactly, but it means you should check before subscribing rather than trusting any published figure, including ours.
No OCR is the real limitation. It's been the top request for years and it still isn't there, while the company shipped a flashcard generator and an AI detector. That tells you something about where development attention went in 2026.
For students and professionals who occasionally need to find something inside a long document: the free tier is genuinely useful and the paid tier is cheap enough not to agonize over. For anyone doing sustained research across multiple sources: NotebookLM is better and free, and the comparison isn't close. For anyone working with scanned material: this tool cannot help you, and no amount of paying changes that.
Note: AIVario earns no commission from ChatPDF. This page is based on published documentation, app store listings and independent reviews rather than sustained use.
Best for: Students reading academic papers, professionals checking contracts and reports, anyone wanting page-cited answers from a single document, users who value zero setup
Not ideal for: Scanned or image-based documents, cross-document research and synthesis, high-stakes accuracy without verification, large image-heavy files
Bottom line: Excellent at one narrow thing, expanded in 2026 into a lot of adjacent things, still missing the one fix everyone asked for. Free tier worth using; check the price before paying.
- NotebookLM — better multi-document research, and free
- Elicit — purpose-built for academic literature review
- Consensus — answers questions across published research
- Claude — larger context and better reasoning over uploaded documents
- Hebbia — enterprise-grade document analysis for finance and legal
- ScholarAI — academic paper search and analysis
- Glasp — highlighting and knowledge capture across sources
Frequently Asked Questions about ChatPDF
What does ChatPDF cost?
This is genuinely hard to state precisely, and anyone giving you one confident number is oversimplifying. Credible sources published within months of each other report paid pricing at $5, $9.99, $14.99 and $19.99 a month. The most detailed recent breakdowns describe a two-tier paid structure — roughly $5 for Plus and $15 for Max — while Apple's App Store listing shows a monthly subscription around $14.99 and an annual plan near $89.99. The most likely explanation is that web and mobile in-app prices differ and that the tier structure changed during 2026. Check the web checkout and the app store separately before subscribing.
Is the free tier usable?
For light work, yes. You get two to three documents a day, around 50 questions each, and a page cap in the region of 120 pages per document. For a student reading a few papers a week or someone checking a contract occasionally, that covers the need. It stops working the moment you're screening literature at volume or handling long documents — a 300-page report simply won't upload on free. As an evaluation tier it's genuinely good, which is more than most tools in this category manage.
What changed in 2026?
The product expanded considerably beyond its name. Page limits on the higher paid tier rose to around 2,000 pages, and support was added for PowerPoint, ePub, Word, Markdown and plain text alongside PDF. Multi-file folders arrived, addressing the long-standing complaint that you could only chat with one document at a time. The company also added YouTube video chat, flashcard generation, slide creation, an AI writer, an AI detector and a desktop app. Whether that breadth improves the product or dilutes what made it good is a fair question.
Does it work on scanned documents?
No, and this is the single biggest limitation. ChatPDF has no OCR, so a scanned page is an image as far as the tool is concerned and you'll get garbage responses or nothing at all. If your documents are photographed contracts, scanned archival material or anything that hasn't been through text recognition, this tool cannot help you. Run them through an OCR step first, or use something that handles it natively.
How accurate is it?
Good on straightforward text documents, less reliable as complexity increases. The page citations are the redeeming feature — every answer points at where it came from, so verification takes seconds rather than requiring you to re-read the document. Users do report inconsistencies on long or structurally complex documents, and hallucinated citations turn up in enough accounts to take seriously. The practical rule: use it to find things, then read the cited page yourself before relying on anything that matters.
ChatPDF or NotebookLM?
NotebookLM handles multiple documents simultaneously, does genuine cross-source synthesis, and is free. On capability it's the stronger tool for research work involving several sources. ChatPDF's argument is interface — it's faster to drag a single file in and start asking, with no notebook to set up. If you're regularly working across a body of literature, NotebookLM is the better choice and the price makes it hard to argue with. If you want one document interrogated right now, ChatPDF is quicker.
What are the file limits?
Reported limits are around 2,000 pages and 32MB per file on the higher paid tier, with the free tier capped near 120 pages. The 32MB ceiling catches people out more than the page count does, because image-heavy PDFs hit it well before they hit any page limit. Users also report that failed uploads sometimes still count against daily allowances, which is worth knowing if you're working near the free tier's limits.
