Tools
Four open-source tools. One checks bank statements against the bank's own balances; three come out of research into characters that pass for one another.
Does it add up?
Drop in PDF bank statements. tothepenny reads every line, checks it against the bank's own balances, and hands you a spreadsheet, with gaps between statements and transfers between accounts marked. It runs in the tab, so the statements never leave your computer.
Which characters pass for one another?
Measures how alike Unicode characters look, font by font, at the size and baseline position each glyph has in running text. RaySpace casts rays through each glyph's outline and compares what they hit, within one font and across fonts, the way a browser shows a rare-script character next to Latin. Every letter and digit in 322 fonts is compared, and the lookalikes of ASCII letters are checked in place, in common fonts at the size people read them. Released as a versioned dataset.
Is this name really free?
When yourapp.com/:slug could be a user, an organisation or a reserved route, namespace-guard checks all of them in one call. Uniqueness across tables in parallel, NFKC normalisation, homoglyph detection from Unicode's latest confusables data plus 2,300 measured lookalike pairs, profanity filtering and suggestions when a name is taken.
Used by agent-sanitizerdeps.dev
Who is registering your name?
Enter a domain and get back its lookalikes, scored by how alike they look, with the ones already registered flagged. Each variant is checked against what the registry actually accepts, for 1,403 TLDs, and a registered lookalike is reported even where today's rules would refuse it. Built on confusable-vision's measurements and namespace-guard's matching, running as a Cloudflare Worker with DNS lookups over HTTPS.
How they connect
confusable-vision measures which characters look alike. namespace-guard ships those measurements as its confusable maps, and d0ma1n uses namespace-guard to find lookalike domains. The research posts tell the story in order.
