← Board

lightseekorg / tokenspeed

A speed-of-light LLM inference engine, TokenSpeed separates control and execution planes, highlighted by the PyTorch Foundation and backed by NVIDIA and AMD.

Python +30 today — week 2.6k total
Maintainer organization
GitHub account
@lightseekorg id 238925110
pump.fun fee address
64QvJMY2XXTsFHnMgwU6rGmNVkFUshbHQPkPH9n2T7GM
Waiting now
0 SOL unclaimed, from any coin
Claimed to date
0 SOL
Repository
GitHub description
TokenSpeed is a speed-of-light LLM inference engine.
Language
Python
License
MIT
Created
2026-05-06
Last push
2026-10-16
Stars · forks
2609 · 345
Listed from
GitHub trending
Summary
One sentence written from the README by a local model (gemma3), checked for length and tone, cached once.