TokenTrim
Trim your AI prompts. Same answer, 40% fewer tokens.
15 followers
Trim your AI prompts. Same answer, 40% fewer tokens.
15 followers
Free Chrome extension that compresses your prompt before it reaches ChatGPT, Claude or Gemini. One click, see the diff, send. Runs locally, stores nothing.
Free
Launch Team / Built With
Atlassian LoomShow, don't just tell with Loom
Promoted
This is a clever little optimization.
I’m curious about one thing: does TokenTrim learn from the way a user writes over time, or are the optimization rules the same for everyone? A personalized approach could make the suggestions much more useful without changing the core workflow.
@rohithreddy Thank you! Honest answer: the rules are the same for everyone today. Everything runs locally with a fixed rule set, and nothing about how you write is stored, which was a deliberate trade-off for privacy.
That said, you've put your finger on the most interesting next step. The version I'm sketching is a light one: when you hit "View changes" and put a word back, the extension remembers that phrase and stops removing it for you. Same workflow, no data leaving the browser, and it adapts to your voice over time.
Curious whether that's the kind of personalization you had in mind, or something closer to learning your whole style?
@rohithreddy Thank you! Great question, and the answer is changing this week because of it. I've just submitted version 0.2 to the Chrome Web Store (hoping for approval soon): open "View changes", click any word TokenTrim removed, and it never removes it again. Like saying "please"? One click and it stays.
It's personalization by correction rather than by profiling: TokenTrim learns only from what you explicitly tell it, and the list lives only in your browser, so nothing about how you write is stored or sent anywhere.
Would love to hear whether that matches what you had in mind, or whether you were picturing something closer to learning your whole style. I'll drop a note here once it's live.