RC RANDOM CHAOS

llm-echo 0.5 refines token usage tracking for testing

· via Simon Willison

Original source

llm-echo 0.5

Simon Willison →

Simon Willison has released llm-echo 0.5, a debug plugin that simulates an language model by echoing back prompts without making actual model calls. It serves as a testing tool for development workflows, avoiding the need to invoke a real LLM during testing.

The new version enhances token usage tracking to account for tool calls, system prompts, and reasoning tokens—previously omitted from the count. This provides more accurate token consumption estimates when developers test and measure interactions during development.

Read the full article

Continue reading at Simon Willison →

This is an AI-generated summary. Read the original for the full story.