llmcord

Why we included this project

Teams that already work in Discord can use this bot as a shared way to try out different models without leaving the server. You @ the bot to start a conversation and reply to keep it going; the reply chain itself is the history, stored in Discord, so there's no database to set up or maintain. It works with any OpenAI-compatible endpoint, so the same bot can reach hosted providers like OpenRouter or OpenAI and local servers like Ollama or vLLM, and a /model command switches between them mid-chat. It also passes along image and text-file attachments when the model supports them, and the config hot-reloads, so changing the system prompt or message limits doesn't mean restarting the bot. For small teams that want a low-friction, collaborative way to test and compare models, this saves building a custom frontend.