1. Load the model Not loaded
The model downloads once (cached by your browser) and then runs on your own hardware using WebGPU. First load pulls a few hundred MB, later visits are instant.
Preparing...
2. LinkedIn headline optimizer
Paste your current headline and role. Get three sharper rewrites tuned for recruiter search and click-through.
Load the model (or set a key) to enable this tool.
3. Run any prompt
This is what the site's "copy this prompt" buttons should have been all along. Type a prompt, get a real answer, streamed live.
Load the model (or set a key) to enable this tool.
⚙️ Bring your own key (optional, still free on Groq/OpenRouter)
Prefer a bigger cloud model or no WebGPU? Paste your own API key from an OpenAI-compatible provider. It is stored only in this browser (localStorage) and calls the provider directly. The site owner pays nothing.
Why this is free
- Hosting: GitHub Pages, static, $0.
- AI: the model runs on your device via WebGPU, so there is no server bill no matter how many people use it.
- Optional key mode: if you bring your own key, you talk to the provider directly. The owner never sees it and never pays.