HuggingChat

HuggingChat is a free, open-source chat interface for accessing and switching between multiple AI models. It offers an Omni mode that automatically selects the best model for your query, or you can manually choose from available open-source models. The chat interface supports fast inference, with examples like Gemma-4-31B running at 700 tokens per second via Cerebras. Users can start new chats and access model settings. The service is free and open-source. Ideal for developers and solopreneurs who want to experiment with different open-source models without cost, for prototyping, research, or casual use.