🤖
Two brains, now connected. The words Clud actually
says are written by a real
in-browser LLM (
@mlc-ai/web-llm on WebGPU) running locally
in your browser — no server, no API key.
This page is the other brain: the shared
hive-mind memory of keywords and topics that users have fed Clud over time. Its top
keywords and topics are now injected into the LLM's prompt, so Clud can casually reference what the
project has actually been about. See
the about page for the full breakdown.
<naughty word> — explicit content
<totally not a secret> — possible credential
📝 Collected Inputs (newest first)
⚠️ DISCLAIMER: Clud's brain is backed by Cosmos DB — data persists across deploys and cold starts.
Inputs are sanitized (passwords, secrets, and explicit content are scrubbed) but keywords and topics are stored indefinitely.
No actual AI is harmed or trained. This is purely for vibes.
If you typed something embarrassing: the sanitizer probably caught it, but Clud makes no promises.