Discord Ollama Bot
A self-hosted Discord bot that connects chat to local Ollama models, with streamed replies, conversation history, private threads and Docker support.
Discord Ollama Bot connects locally hosted language models to Discord, so a community can use an AI assistant inside its existing chat server. It's for server owners who want to choose and host the models themselves while keeping Discord as the interface. Model inference runs on your hardware, but conversations still pass through Discord, so this isn't a fully offline chat app.
The bot streams replies as the model generates them and keeps conversation history for ongoing exchanges. Private threads give users a separate place to talk with the assistant. It also supports model management and configurable server controls, giving the person hosting it control over the models available to users and how the bot operates in the server. Users interact through Discord slash commands.
Ollama is the main backend. The bot can also connect to llmman, which provides the Ollama API. You can host it on a local machine or a server, with Docker support for both; the project recommends NVIDIA GPUs and discourages local-machine use. The distinction matters for anyone looking for a desktop chat app: this tool connects a model host to a Discord community.
The bot is written in TypeScript and carries the Creative Commons Attribution 4.0 International (CC-BY-4.0) license.
