It reads your whole library the way its author would.
Ask a question and Chunkless finds the clause in one file, the exception buried in another and the memo that quietly overrode both. It answers with the exact section and page, tells you when your documents disagree, and says so plainly when they don't say at all.
Your seat in Chunkless is ready - welcome. Your invite code is already in this link: set a password and your workspace is ready.
Set up my access →Finance Memo 2025-07
Asset Register
Procurement Policy
Travel & Expense Policy
Yes, with two conditions. It is within a director's $75,000 limit (Finance Memo 2025-07), and software is operating spend, never capital, so the CFO rule for capital items does not apply. It still needs three written quotes (§5.1) from an approved vendor (§5.2).
Search finds a paragraph. Your question lives across three documents.
Most AI search cuts your documents into chunks, grabs the few that sound like your question and writes something fluent on top. It sounds right. Then someone finds the exception on page 40.
Chunkless doesn't cut documents into pieces. It keeps each one whole — its sections, tables, figures and pages — and follows what they point to, the way a person who wrote them would.
What it does
Built for the questions nobody can answer from one document alone.
Links anything to anythingAcross files, departments, years
It learns your documents' vocabulary: the same thing under three names, what refers to what, which version replaced which. Then it can connect something in volume 1 to volume 20 on its own.
Answers with its sourcesSection · page · figure
Every part of the answer traces back to where it came from. Open the source, see the passage, check it yourself. Nothing is stated that the documents don't support.
Tells you what it can't proveComplete · conditional · missing · conflicting
It doesn't paper over gaps. Each answer says whether the documents settle it, settle it only under a condition, point to something you haven't uploaded, or contradict each other.
Explains why something is going wrongRoot cause across tools and processes
Ask “why does this keep failing?” and it traces the chain through runbooks, configs and procedures to the step or rule that's actually causing it. It shows the chain, not a guess.
Reads what text search can'tFigures · tables · scans · spreadsheets
Every diagram and screenshot is read and described when you upload it, tables keep their rows and columns, and scanned pages are OCR'd. A 300-step manual answers with each step's picture.
Thinks with you, not just for youResearch mode
For the long questions: give it the papers or the specs and ask for a design, a theory, a plan. It pieces an answer together from what the documents prove and marks what is still open.
Organized the way your company isOrganization → wing → department → case
Put HR, Finance, Legal or Cardiology in their own departments, with folders and cases inside. Ask one department, a selection, or the whole organization.
Takes documents as they arePDF · Word · PowerPoint · Excel · scans
Drag in files up to 500 MB, or connect an S3 bucket. Connectors for Google Drive, OneDrive, Dropbox and Box are on the way.
Private by constructionIsolation · encryption · your model
Each organization is isolated at the database's row level, so one tester can never read another's documents. Connector keys are encrypted at rest. You can bring your own model key.
Every answer comes with a verdict.
A good expert tells you how sure they are and why. Chunkless works out that verdict from your documents instead of guessing it.
The documents settle it.
Everything the answer needs is in what you uploaded, and each piece is cited.
“Leave carries over 5 days.” — HR Handbook §6.2
True, if…
The answer depends on something your question didn't say, and it tells you exactly what.
Yes, if the vendor is already on the approved list (§2.4).
A document points elsewhere.
The rule refers to a schedule, form or policy that isn't in the library yet, and it names it.
§9 depends on “Schedule B”, which isn't uploaded.
Two sources disagree.
Both are in force and they can't both be right. It shows you both, and who decides.
Memo 2025-07 vs Policy §4.2: the stricter applies until Finance rules (Delegation §3).
Heavy reading once. Sharp answers after that.
It works like the expert you'd want to ask. It reads everything carefully up front, remembers how it all fits together, and when you ask, it goes back to the sources and checks before it speaks.
Bring the library
Upload files, or point it at a bucket. Put them in departments and cases the way your team thinks about them.
- PDF, Word, slides, sheets, scans
- up to 500 MB a file
It reads everything
Sections, tables and figures are kept in place, with every rule, limit, definition and cross-reference noted.
- every figure described
- names linked by meaning
- what overrides what
You ask anything
A quick lookup, a “can I…?”, a “why does this keep breaking?”, or a design built from thirty papers.
- one department or all of them
- follow-up questions keep context
It investigates
It opens the sources, looks for evidence against its first answer and checks every claim before you see it.
- cited to section and page
- verdict: complete → conflicting
- the reasoning trail, on request
Match the effort to the question. You pick, or it picks for you.
Quick
A fact, a number, a definition, in seconds.
Full
Reads until each part is found or shown absent, then checks every claim. Minutes, not seconds.
Auto
The engine picks: quick for a one-fact lookup, full for everything else.
Research
Long-running reasoning: designs, theories, plans.
Anyone who inherits a pile of documents and a deadline.
Can I approve this, and who says so?
A new hire gets the answer a ten-year veteran would give, with the policy, the memo that changed it and the exception that applies.
Why does the export fail every Monday?
It follows the chain through runbooks, configs and procedures to the step that's actually causing it, and shows the chain.
From these papers, what should we build?
It pieces a design together from what each source proves, says where they conflict and marks what's still an open question.
We're letting people in a few at a time.
This is a working test build, not a demo video. We invite a small group at a time so everyone gets a fast, reliable engine and a direct line to the people building it.
- CostFree during early access.
- Your inviteA personal code emailed to you. It works once, for your email address only, and lasts 14 days.
- Your access7 days from the day you join. Genuine feedback through the Feedback button in the app earns extra days.
- Your dataYour organization is walled off from every other tester. Please don't upload anything you aren't allowed to share with a test system.
- What we askTry it on real questions and tell us where it's wrong. That feedback shapes what ships next.
You're in line.
We read every request. If yours is a fit for this round, we'll email you a personal invite code. It will only work with the address you gave us.
Nothing arrived after a while? Check your spam folder for mail from Chunkless.
Before you ask.
How is this different from ChatGPT with file upload?
Those tools read a handful of passages that sound like your question. Chunkless reads your whole library up front, and learns what refers to what, which rule overrides which and where each figure belongs. When you ask, it investigates and checks every claim against its source. So it can say “these two documents disagree” or “the answer depends on a schedule you haven't uploaded” instead of writing something that merely sounds right.
How do invites work?
Joining puts you on the list with your reason. We read every request and send invites by hand, each with a personal code tied to your email address. The code works once and expires after 14 days. The link in the email takes you straight to setting a password, and then you're in your own workspace for 7 days. The Feedback button at the top of the app reaches us directly: report what's wrong, suggest what's missing, and genuine feedback earns extra days.
Where do my documents go?
They're stored encrypted on AWS in the US. Each organization is isolated at the database's row level, so no other tester can see your documents. To answer questions, the engine sends the relevant passages to a language model (OpenAI by default), or you can bring your own model key. During early access, please only upload documents you're allowed to share with a test system.
What kind of documents work best?
Collections where the answer is spread out: policies and their amendments, SOPs and manuals, contracts and schedules, research papers, clinical protocols, engineering runbooks. It handles PDFs (including scans), Word, PowerPoint and Excel.
Why is access limited?
Reading documents deeply takes real compute. We'd rather give a small group a fast, dependable engine than give everyone a slow one. We open it to more people as we scale.