https://cloudposse.com logo
Join Slack
Powered by
<https://github.com/chopratejas/headroom>
# random
h
forgot to say this is not my project
e
About
Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95% fewer tokens, same answers. Library, proxy, MCP server.
I remember there was some talk about this; there was another project that turned context/prompts into caveman speak
...To reduce tokens.
My understanding is it reduces the quality of the outputs
Holy cow that caveman project got 67K stars!
h
wow, that works then, 20$ plan can work like 5x max
r
Caveman was novel but it was quickly replaced by rtk and then headroom (created by netflix and wiz) uses rtk under the hood and applies to a lot more
👀 1
They have a good comparison on the readme of their doc but caveman isnt included
I replaced caveman with rtk a couple months ago and the gains are amazing
h
wow wow
great to know, 5x max is enough for me, may need them if there is a new project other than openclaw or hermes (more secure than them)
2 Views
Open in Slack
PreviousNext
Powered by