Browse all MCP servers by davz33.
by Davz33
This server implements speculative decoding with a local LLM, prioritizing local model responses and falling back to other models when necessary.