AI Tools

What is sent to an AI, and what is not

The generative tools send the text you enter, together with your chosen options, to a hosted open-weight language model to produce the result; that is the one case on this site where your input leaves your browser. The request carries no account or name, nothing is used for training, and the result is returned to the page and not kept. Three tools in this category do not use a language model at all: speech-to-text and text-to-speech use your browser's own speech engines, and the AI-content detector scores writing signals locally. Each tool page states which kind it is.

Getting a usable result

• Give the tool context, not just a topic: audience, purpose and anything it must include. A product description from a list of real specifications beats one from a name alone.
• Pick the tone option that matches where the text will be used — the same draft reads differently as a LinkedIn post and as a customer email.
• Treat the output as a first draft. Check facts, names and numbers; a language model writes fluently whether or not it is right.
• For long inputs, summarise or rewrite in sections. Every tool shows its input limit on the page.

Limits and fair use

The AI tools are free and rate-limited per visitor to keep them available to everyone. If a model is temporarily unavailable, the tool falls back to another one automatically and tells you if none responded. Results are not guaranteed to be unique, and the AI-content detector's score is a heuristic that produces false positives — never use it as proof that a person did or did not write something.