I built VeloTools because I was tired of sending client photos, invoices, and PDFs to random SaaS boxes just to shrink a file. “We delete after 24 hours” is not privacy. It is a retention policy you cannot audit.
If there is no upload endpoint, there is nothing to leak, queue, or sell.
What we actually do
Image compression, resize, and convert run on the Canvas and File APIs in your browser. Background removal loads a WebAssembly model on-device. PDF compression uses PDF.js / pdf-lib locally. Invoices and QR codes are generated client-side. After the page loads, you can go offline and the tools still work.
What that forces us to refuse
- We do not claim medical, legal, or financial advice. Tools change pixels and documents — they do not diagnose or certify.
- We do not promise “instant indexing” or ranking. Sitemap + quality pages are the discovery path; Indexing API is only an accelerator.
- We do not publish thin PSEO clones that only swap a word in the title. If the math, preset, or workflow is identical, the page does not ship.
- We do not make AI chat the face of the site. The face is the instrument you operate with your hands.
Limits I will tell you up front
Browser RAM is the real ceiling — a huge scanned PDF can stress a phone. WASM models need a first download. Some RAW cameras only expose an embedded JPEG preview, not a full demosaic. Saying that out loud is part of trust.
Why this note exists
Search platforms put young sites in a sandbox when they look like faceless content farms. Lab Notes are not a blog for keywords. They are a record of how the product is built, by whom, and what we refuse. If you only need the tool, skip the essay and open it.