Assistant
Ask about a regex, a broken config, a JWT that will not validate, or a cron expression you cannot read. It knows the tools on this site and will point you at the right one.
Start with one of these, or ask anything.
What this is, and what it is not
The assistant runs on free models routed through OpenRouter. There is no account, no sign-up and no charge. In exchange it is not as capable as a paid frontier model, it can be slow when the free tiers are busy, and like every language model it will occasionally state something wrong with complete confidence. Check anything that matters.
Why it lives on its own page
Every other tool on this site processes your input locally, which is the entire reason the site exists. An assistant cannot work that way — the model is not on your machine. Rather than bury that difference, it gets its own page and a notice at the top, and there is no chat widget floating over the tools. When you are using the hash generator or the password generator, nothing is being transmitted, and that should never be ambiguous.
What not to paste
The same advice as for any chat service, and for the same reasons set out in why you should never paste a secret into an online tool: no credentials, no private keys, no customer data, no production tokens. If you want to inspect a real JWT, use the decoder — it never sends anything anywhere.
Limits
There is an hourly message allowance per visitor, because the free model tiers have finite capacity and an open endpoint gets used as a free API within days. If you hit it, the tools themselves remain entirely unrestricted — they always will, because they cost nothing to run.
Your conversation stays on this device
The transcript is kept in your browser's local storage so it survives a reload. It is never stored on our side. Clearing site data, or pressing New chat, removes it permanently.