Semble Hooks — semantic code search injected into every Codex prompt #24451
benediktkraus
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Every prompt you type gets the 5 most relevant code chunks from your project — automatically.
How: Semble indexes your project (<1s, CPU-only, no API key). On every UserPromptSubmit, the hook calls
semble search, reranks results (definition boost, noise penalty), and injects them as<relevant-code>into the prompt.Result: 98% fewer tokens than grep+read. Your AI sees the relevant code BEFORE it starts working.
Also works with Claude Code and Gemini CLI.
git clone https://github.com/benediktkraus/semble-hooks.git cd semble-hooks ./install.sh codexCompanion to DREVIHO (persistent memory hooks).
Repo: https://github.com/benediktkraus/semble-hooks
Beta Was this translation helpful? Give feedback.
All reactions