TEMPLATE

Token Diet (by uniqaiapp)

👁 0

Token diet for LLM apps and coding agents: measure where input tokens go (system prompt, history, tool output, RAG), then trim with stable prefixes, history windows and compact tool results. Claude Code skill + AGENTS.md/Cursor rules + a stdlib Python token-budget report. Numbers

RELATED TEMPLATES

COMMENTS

or sign in to comment under your account

No comments yet — be the first.