
Forge Mcp Server
io.github.RightNow-AIPublished by GitHub account RightNow-AIv1.0.2Updated Oct 24, 2026
Turn PyTorch into fast CUDA/Triton kernels on real datacenter GPUs with up to 14x speedup.
Overview
A local MCP server that lets an assistant turn PyTorch code into fast CUDA/Triton kernels on real datacenter GPUs, with claimed speedups up to 14x.
- What you can do
- Convert PyTorch code into CUDA or Triton kernels for datacenter GPUs, aiming for faster execution than the original PyTorch implementation. The registry description advertises speedups of up to 14x, but no tool list, instructions or README are provided, so the exact operations the assistant can invoke are not documented here.
- When to use it
- Consider it when you have PyTorch workloads you want to accelerate by generating custom CUDA or Triton kernels, and you have access to datacenter GPUs to run them on.
- Requirements
- Runs as a local process on the user's machine, installed from the npm package @rightnow/forge-mcp-server over stdio. It is desktop only and not web executable. No authentication, environment variables or headers are declared. Access to real datacenter GPUs is implied by the description.
Installation
In SourceWeft
- Open Forge Mcp Server in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
This repository has no README.
Tools
0Version history
1- v1.0.2LatestSep 16, 2026
Score
25% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
Declare a GitHub repository with a README, or serve resource_documentation in the OAuth protected-resource metadata.
- License10 pts
Declare a license in the repository or the registry entry.