llama-swap ghcr.io/mostlygeek/llama-swap:unified-cuda ghcr.io/mostlygeek/llama-swap unified-cuda Latest stable release bridge http://[IP]:[PORT:8080]/ false https://github.com/mostlygeek/llama-swap/issues https://github.com/mostlygeek/llama-swap Run multiple generative AI models on your machine and hot-swap between them on demand. llama-swap works with any OpenAI and Anthropic API compatible server and is used by thousands of people to power their local AI workflows. False AI: Productivity: Tools: Other: Status:Stable ai models llama ollama swap change multiple openai anthropic api server https://raw.githubusercontent.com/nwithan8/unraid_templates/master/images/llama-swap-icon.png https://raw.githubusercontent.com/nwithan8/unraid_templates/main/templates/llama_swap.xml https://github.com/nwithan8 ### 2026-06-28 Initial release Expects a `config.yaml` file at the Config File path BEFORE starting the container. --runtime=nvidia --gpus all 9292 /mnt/user/appdata/llama_swap/config.yaml /mnt/user/appdata/llama_swap/models