compare_models

Instructions

Implementation Reference

• src/openrouter_mcp/server.py:198-230 (handler)

  The @mcp.tool() decorated function that implements the 'compare_models' tool logic. It fetches models, matches given comma-separated model IDs (exact or fuzzy), and renders a side-by-side comparison table with Context, Input $/1M, Output $/1M, and Tools support columns.

  @mcp.tool()
  def compare_models(model_ids: str) -> str:
      """Compare multiple models side by side.
  
      Args:
          model_ids: Comma-separated model IDs
      """
      models = fetch_models()
      ids = [x.strip() for x in model_ids.split(",")]
      selected = []
      for mid in ids:
          for m in models:
              if m["id"] == mid or mid.lower() in m["id"].lower():
                  selected.append(m)
                  break
      if not selected:
          return "No matching models found."
  
      names = [m["id"] for m in selected]
      max_len = max(len(n) for n in names) if names else 10
  
      def row(label, fn):
          vals = " | ".join(str(fn(m)).rjust(max_len) for m in selected)
          return f"| {label} | {vals} |"
  
      lines = ["# Model Comparison\n"]
      lines.append("| Metric | " + " | ".join(n.rjust(max_len) for n in names) + " |")
      lines.append("|--------|" + "|".join("-" * (max_len + 2) for _ in names) + "|")
      lines.append(row("Context", lambda m: m.get("context_length", "?")))
      lines.append(row("Input $/1M", lambda m: _price_str(float(m.get("pricing",{}).get("prompt",0)))))
      lines.append(row("Output $/1M", lambda m: _price_str(float(m.get("pricing",{}).get("completion",0)))))
      lines.append(row("Tools", lambda m: "✅" if "tools" in m.get("supported_parameters",[]) else "❌"))
      return "\n".join(lines)

• src/openrouter_mcp/server.py:198-199 (registration)

  The @mcp.tool() decorator registers 'compare_models' as an MCP tool on the FastMCP server instance.

  @mcp.tool()
  def compare_models(model_ids: str) -> str:

• src/openrouter_mcp/server.py:55-58 (helper)

  Helper function _price_str converts per-token price to a human-readable string like '$2.50/1M tok' or 'free', used by compare_models when displaying pricing columns.

  def _price_str(per_token: float) -> str:
      if per_token <= 0:
          return "free"
      return f"${per_token * 1_000_000:.2f}/1M tok"

• src/openrouter_mcp/server.py:34-52 (helper)

  The fetch_models helper fetches the OpenRouter model list (with caching), called by compare_models to retrieve model data.

  def fetch_models(force=False) -> list[dict]:
      """Fetch model list from OpenRouter with caching."""
      now = time.time()
      if _cache["data"] is not None and (now - _cache["ts"]) < CACHE_TTL and not force:
          return _cache["data"]
  
      headers = {"Accept": "application/json"}
      if OR_API_KEY:
          headers["Authorization"] = f"Bearer {OR_API_KEY}"
  
      req = Request(OR_MODELS_URL, headers=headers)
      try:
          with urlopen(req, timeout=30) as resp:
              body = json.loads(resp.read())
              _cache["data"] = body.get("data", [])
              _cache["ts"] = now
              return _cache["data"]
      except URLError as e:
          raise RuntimeError(f"Failed to fetch OpenRouter models: {e}")

• src/openrouter_mcp/__init__.py:5-7 (registration)

  The compare_models function is exported from the package via the __init__.py __all__ list and direct import.

  from .server import main, fetch_models, list_models, get_model, search_models, compare_models, refresh_cache
  
  __all__ = ["main", "fetch_models", "list_models", "get_model", "search_models", "compare_models", "refresh_cache"]