Forge MCP Server transforms PyTorch models into optimized CUDA/Triton kernels using multiple AI agents for faster inference. It connects MCP-compatible AI coding agents to optimize and benchmark code on real datacenter GPUs.
From the registry: Turn PyTorch into fast CUDA/Triton kernels on real datacenter GPUs with up to 14x speedup.

Please install the `forge-mcp-server` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **RightNow AI account** — Authenticate via browser-based OAuth (no API key needed). Call forge_auth to sign in. (https://dashboard.rightnowai.co)
Canonical MCP server config (stdio transport):
- command: `npx`
- args: ["-y","@rightnow/forge-mcp-server"]
Note: OAuth-based auth — no env vars required. Turns PyTorch into CUDA/Triton kernels via swarm agents on datacenter GPUs. 1 credit per optimization.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.AI Agents Framework with Self Reflection and MCP support