Every day, people generate more than 10,000 free downloads with DocuPipe. Converted spreadsheets, redacted PDFs, split documents, searchable scans, consolidated trackers - produced by our engine, downloaded mostly by people who never signed up, never entered a card, and will never pay us a cent.

I want to be clear about how we feel about that number: delighted. We watch it the way other companies watch revenue. And we just shipped four more tools to push it higher.

DocuPipe's growth has always worked one way: deliver value long before we extract anything. The platform has a free tier that runs real extractions. The API works before a sales call. And the free tools take that idea to its logical end - the entire value, delivered in the browser, in under a minute, to someone we know nothing about.

There's no trick hiding in that sentence. No watermark held hostage, no "your file is ready, create an account to download it," no three-conversions-then-paywall meter. Drop the document, get the result, leave. The single-document experience is free and complete, on the same certified infrastructure as everything else we run.

Why does that make business sense? Because document work is never actually a single document. The person converting one loss run today works at a desk where five thousand arrive every month. The ops manager pasting eight order emails into a tracker has a mailbox that refills every morning. When the one-document version of the job costs nothing and works, the ten-thousand-document version of the job has an obvious address. Some fraction of free users have that industrial volume, and they graduate when they're ready. Everyone else got a free spreadsheet, and that's a fine outcome too - they'll remember where it came from, and a few of them will write about it or link to it. That's the whole model. We'd rather earn distribution by being useful ten thousand times a day than rent it.

The internet is full of converter farms - hundreds of near-identical "X to Y" pages wrapping the same generic engine. We deliberately went the other way: fewer tools, each built as a real product for one professional job, in that profession's vocabulary, with its own interface and its limits stated plainly.

The lineup so far: PDF to Excel, Loss Run to Excel for insurance claims histories, the Rent Roll Normalizer with its totals tie-out, the Auto Redactor built around a review screen, Scanned PDF Search for piles of image-only scans, the Document Splitter, the COI Tracker, Quote Comparison, and a Hebrew OCR page written entirely in Hebrew.

This week, four more - one for each way a table gets trapped outside a PDF:

Every tool ships with fictional sample documents so you can see the whole flow without uploading anything of your own, and every page states its caps plainly: file size, page count, daily runs. When a tool can't do something - fetch a URL, read an .eml file - the page says so instead of pretending.

No signup. A generous daily allowance per network. Files encrypted in transit and at rest, processed on SOC 2 and ISO 27001 certified, HIPAA compliant infrastructure, and never used to train models - the same guarantees our paying customers get, because it's literally the same engine. The tools aren't a stripped-down demo build; they're the platform, pointed at one job at a time.

That's also the pitch for what we sell. If a free page can turn a tilted phone photo into a workbook in under a minute, that's the extraction quality you're buying when you put DocuPipe behind your document pipeline - at whatever volume your desk actually deals with, with schemas you control.

Ten thousand downloads a day, and climbing. We'll keep building these. If there's a document job you keep doing by hand and wish were on the shelf, tell us - the shelf is at www.docupipe.ai/tools.