wtf is discord-voice-status-summary?

This is a proof of concept Discord bot that listens to a voice channel call, summarizes what people are saying, and shows that summary as the voice channel's status text. It is written in TypeScript using discord.js and the AI SDK. The bot itself does not do the speech to text or the summarizing. Instead it passes recorded audio to a separate transcription server called speaches, which offers an OpenAI compatible API, and then sends that transcript to LM Studio, a local tool for running language models, to produce a short summary. That summary is what appears on the voice channel status. The README includes an example Docker Compose file for running speaches with an NVIDIA GPU, which needs the nvidia container toolkit installed on the host machine. To use it, you first create a bot application in the Discord developer portal and invite it with permission to connect to voice channels and set the voice channel status. Then you start an LM Studio server with a loaded model, copy an example environment file and fill in the Discord token plus the URLs and model names for the transcription and summary servers, and start everything with Docker Compose. Once a user joins a voice channel and runs a join command, the bot periodically updates the status with a fresh summary, on an interval you configure, until someone runs a leave command. Configuration is handled through environment variables that are validated on startup, covering the Discord token, the transcription server URL and model, the transcription language, the LM Studio URL and model, and how often the summary refreshes. For development, the project uses pnpm to install dependencies, run the bot directly, or build it. Pushing to the main branch triggers a GitHub Actions workflow that builds and publishes a Docker image for the project automatically.