WhichModel
- whichmodel/whichmodel-mcp
WhichModel is an MCP server that delivers cost-optimized LLM routing recommendations for autonomous AI agents over HTTP.
About
WhichModel is a hosted MCP server that returns cost-optimized large language model routing recommendations tailored to autonomous AI agents. developers burning budget on default frontier models can register the streamable-http endpoint so their agent picks cheaper-fit models per task without maintaining a spreadsheet of provider pricing. It sits in the build phase as agent-tooling: use it when designing agent loops, sub-agent delegation, or high-volume automation where margin matters as much as quality. Connect via https://whichmodel.dev/mcp in a compatible client; pair recommendations with your own evals because routing is economic guidance, not a guarantee of output quality. Ideal for agents shipping on Claude Code, Cursor, or Codex stacks with variable step costs.
- WhichModel remote MCP at https://whichmodel.dev/mcp
- Cost-optimized LLM routing recommendations for autonomous agents
- Streamable-http hosted server v1.0.0
- Public site https://whichmodel.dev for context alongside MCP
WhichModel by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
claude mcp add --transport http whichmodel https://whichmodel.dev/mcpAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| Transport | HTTP |
|---|---|
| Auth | None |
| Repository | whichmodel/whichmodel-mcp ↗ |
What it does
Route autonomous agent workloads to cost-efficient LLMs using WhichModel recommendations over remote MCP.
Who is it for?
Best when you're running high-volume or multi-agent automations and need MCP-accessible cost routing without building your own router.
Skip if: Single-shot chat with one fixed model, or teams that ignore cost and only optimize for maximum capability per call.
What you get
Your agent can query WhichModel for cheaper-fit model choices while you ship and iterate multi-step autonomous workflows.
- Agent-callable cost-optimized model routing recommendations
- Hosted MCP integration without local npm stdio install
- Reference alignment to WhichModel v1.0.0 remote server
By the numbers
- Server version 1.0.0
- Remote MCP URL https://whichmodel.dev/mcp
- Transport type streamable-http
Recommended MCP Servers
0Latency MemoryPersistent memory layer for AI agents.
0nMCP — Universal AI API OrchestratorUniversal AI API Orchestrator — 1,554 tools, 96 services.
0xHumans Protocol MCPMCP for AI agents: financing, skills, lending on Base1k Patient Mcp1k patient MCP server
1trippulse1trip PULSE: 21-tool AI travel planner.
3D AI Agent Avatar3D AI Agent Avatar — render any GLB, give it a Solana wallet, a voice, and pump.fun powers.89How it compares
LLM cost-routing MCP service, not a prompt library skill or local model runner.
FAQ
Who is WhichModel for?
Developers and agent authors who want autonomous workflows to pick cost-efficient LLMs via MCP instead of hard-coding one provider.
When should I use WhichModel?
Use it while designing or operating agent loops where token spend varies by step and you want routing recommendations on demand.
How do I add WhichModel to my agent?
Configure your MCP client with the remote streamable-http URL https://whichmodel.dev/mcp and verify connectivity to whichmodel.dev.