Embeddings is all you need

Initialising

Embeddings is all you need

No LLM<br>Embedding model only<br>100% offline

Downloading MiniLM embedding model…<br>0%

Voice<br>Embedding<br>Action

100% in-browser · no server · no LLM · Intent classification using a tiny embedding model (MiniLM-L6-v2, 23 MB, WASM) — cosine similarity, not a language model

Click to speak

Transcript

↵ classify<br>✕ clear

🛒 Shopping list

✕ clear

Say "add milk" or "remove bread"…

⏱️ Timer

✕ cancel

⚡ Custom actions

+ Add action

Intent

Confidence

Latency

Example commands — click to trigger with this text

Cosine similarity per intent

Waiting for first command…

Local pipeline · no server · no LLM

Web Speech API<br>Transcript<br>MiniLM embedding (WASM)<br>Cosine similarity<br>DOM action