Tokens and cost per model — and why Arabic costs more.
Developer
15 toolsTools for working with language models and structured text: count tokens and cost per model, build and lint a prompt, turn sample JSON into a schema or tool definition, clean model output, and copy Markdown as rich text. Everything runs on your device; nothing you paste is sent anywhere.
Write a prompt in parts, get it linted and exported.
Sample JSON in, schema and tool definitions out.
Strip fences, fix broken JSON, pull tables to CSV.
Paste model output, get Word- and email-ready text.
Two prompts in, every changed word out — with the token delta per model.
Pattern, flags, test text — matches, groups and export for JS, Python, Go.
Format, minify, sort keys, query a path, convert to YAML or TOML.
Paste a ChatGPT or Claude chat, get clean Markdown or a messages array.
Reading level, rhythm, filler phrases — and the hidden characters you can't see.
Normalise, strip invisibles and tashkeel, fix digits — see every code point.
Token-sized chunks with overlap, Arabic sentence-aware, JSONL out.
Base64, URL, hex, SHA digests and JWT decoding — offline.
WCAG and APCA for colour pairs — or a palette against both Dzzz themes.
Paste curl, get fetch, Node or Python — secrets masked by default.