llmcord
A self-hosted Discord LLM bot that connects to Ollama, LM Studio or cloud APIs, with shared conversations and an MIT open-source license.
llmcord is a self-hosted Discord bot for people who want to share LLM conversations with friends or a community. You can run the Python bot on your own machine or server, including through Docker, and connect it to local models or cloud providers. It's open source under the MIT license.
Discord replies define each conversation's history. People can branch a discussion, pick up someone else's conversation, or carry a branch into a thread. The bot also supports direct messages and recognizes participants by their Discord IDs, so a shared chat can retain who said what. Conversation history stays in Discord, and the bot doesn't need a separate database.
For local inference, it connects to Ollama, LM Studio and vLLM. It also works with OpenRouter, OpenAI, xAI and Google, provided the endpoint supports the OpenAI-compatible chat completions API. Administrators can switch between configured models within Discord.
The bot accepts text and code file attachments, plus images when the selected model supports vision. It streams replies and splits long responses across messages. A custom system prompt lets the host set its behavior, while access controls limit use by Discord user, role or channel.
Local inference doesn't make the chat offline or keep it entirely on your hardware: Discord hosts the messages and requires an internet connection. With a local backend, model processing runs on your machine or server; with a cloud backend, the selected provider receives the conversation context.
