Live demo

Meet a mini Edlan agent

A taste of what a website assistant feels like. This one is a lightweight demo running entirely in your browser. The real thing would be powered by a production LLM.

Edlan Assistant Demo mode, in-browser only
⚡ Real AI mode: load a small open-source model that runs entirely in your browser. One-time download of ~280 MB (be mindful on mobile data); cached for next visits.
Building the magic ✨
Downloading a ~280 MB model. This uses real data, so grab a coffee…

Demo mode uses simple pattern matching: instant, with no download. Real AI mode runs Qwen 2.5 (0.5B), a small open-source language model, entirely in your browser via WebGPU. Nothing you type is sent to any server, the model has no internet access, and no third-party AI service (Claude, OpenAI, etc.) is involved. Small models can make mistakes; a production deployment would use a larger model with stronger safeguards.