Steven Gonsalvez

Software Engineer

Vomit: Clean up Claude 5's token output with a separate LLM

Why CEREBRO kept it

Token optimization via separate LLM.

The text below is an automated extraction of the article at https://github.com/zachahn/vomit, stored verbatim in the public cerebro-vault repository. Copyright remains with the original publisher (github.com).

Vomit converts Claude's token vomit into English by piping it through a local LLM. It's fully local (no telemetry) and has no external dependencies. I wrote a small blog post about this project: https://zachahn.com/posts/1787191554 Disclaimer: - The local LLM can only see what Claude tries to communicate (no access to any actions or files), so it hallucinates a bit - It's pretty slow - This is totally vibe-coded, only tested on Mac - There is a possibility you'll completely miss Claude's message. You can use something like AgentsView to get the original messages, as Vomit does not touch anythi

Community take

Claude 5's strange verbose output requires post-processing; some users switched to Codex to avoid it.

Backlinks

Appeared in 1 briefing

Related

Shares tags: ai/llm-mechanics · cerebro/signal

Also from github.com