Mochi is a configurable LLM-based chatbot for Revolt. It supports per channel models and prompts as well as automatic transient history keeping and basic context awareness.
- Only runs on Python 3.10 right now.
- Requires an Ollama backend.
It's easy to set up and configure:
- Clone the repo.
- In the config directory, copy config.toml.template to config.toml.
- Create a bot in the My Bots section in the Revolt settings.
- Add the bot to a server of your choosing.
- In the toml file: Add your token and adjust the other parameters as desired.
That's it! Just mention the bot in a message and wait for the reply. There are some minor known issues that I may fix over time if there is demand.