Open-source tools for privacy, accessibility, and AI integrity — almost all of them running entirely in your browser: no account, no upload, no tracking, no server. Everything is MIT-licensed.
🧰 Start here — clearbox
Seven everyday tools in one place, in English and Arabic:
| Tool | What it does | |
|---|---|---|
| 🛡️ | Scam message checker | See the red flags in a suspicious SMS, email, or chat message |
| 📄 | Contract & terms checker | Find the clauses worth reading twice before you sign |
| 🔑 | Password checker | The exact reason a password is guessable, not a vague bar |
| 📷 | Photo metadata remover | Strip the GPS location hidden in a photo before sharing it |
| 🎨 | Contrast & color-blindness | Check WCAG contrast and simulate color vision deficiency |
| 📖 | Plain-language simplifier | Make dense text readable, with dyslexia-friendly aids |
| 🧾 | Group expense splitter | Split a shared bill and settle up in the fewest payments |
- citeguard — check whether a citation has been retracted before you use it. Python library + CLI, MCP server for AI agents, and a GitHub Action. Wakefield's fraudulent MMR paper has been cited 1,000+ times since its retraction; this makes that checkable in one command.
- cite-lens · ground-lens — find uncited claims and unused references; check whether each claim is actually supported by its source.
- secretscan — catch leaked API keys and private keys before an agent commits them
- regexguard — explain a regex and flag ReDoS risk, without executing it
- readability — let an agent check how readable its own output is
- sourceward — detect prompt-injection attempts in untrusted text
DialectLens (dialect detection) · MeaningDiff (semantic diff) · TimeLens (timeline extraction) · CoherenceLens (contradiction finder)
Maqam Studio — explore Arabic maqams with true microtones · Zellige Studio — generate Islamic geometric star patterns