go-mcp-gateway

dcotelessa/go-mcp-gateway

★ 0 stars Go AI/LLM Updated today

MCP tool server and local model gateway for llama.cpp — routes LLM completions by task complexity with token budget management and LSP session bridging

View on GitHub → 🔍 Audit Wallet Slippage →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "go-mcp-gateway": {
      "command": "go",
      "args": [
        "run",
        "github.com/dcotelessa/go-mcp-gateway@latest"
      ]
    }
  }
}

README Excerpt

A single Go binary that acts as both an **MCP tool server** and a **local model gateway** — routing LLM completion requests to local llama.cpp instances or remote APIs based on task complexity, with per-session rate limiting, token budget management, and LSP session bridging for code intelligence. Built to solve a real problem: running a multi-agent AI development pipeline on a single GPU workstation where VRAM is the binding constraint and cost control across local and remote models matters.

Tools (5)

multi_filerecoveryscaffoldsingle_filetext_op

go-mcp-gateway

Quick Install

README Excerpt

Tools (5)

Related AI/LLM Servers