Developer

Pasting a production API response into a random online formatter is a habit worth breaking, because that JSON may hold customer data or a token. Everything here processes input in the browser, so it’s safe to use on the real thing.

For everyday chores: the JSON Formatter validates, pretty-prints, sorts keys and minifies. The Regex Tester highlights matches and capture groups live and has a replace mode. The Base64 Encoder handles text and files, and the URL Encoder covers percent-encoding, query strings and Base64URL. Base64 is encoding, not encryption: anyone can decode it.

If you build on LLM APIs, the Token Counter counts tokens, estimates cost and checks context-window fit across GPT, Claude and Gemini models. Its OpenAI counts are exact; the Claude and Gemini figures are estimates, because Anthropic and Google don’t publish a tokenizer that runs in JavaScript, so it counts with the closest OpenAI encoding. The Tokenizer Visualizer colors your text token by token with the GPT-4o and GPT-4 tokenizers, which makes it obvious why code and non-English text cost more than you’d guess. Before you ship, the LLM Cost Calculator turns tokens per request and requests per day into a monthly bill for every priced model, cheapest first.

For the theory, read how GPT tokenizes text.