llm-echo 0.5 released, simulated token usage now includes tool calls
Understand the story
Simon Willison updated his LLM debugging plugin llm-echo. The plugin simulates a "model" that simply echoes back the input prompt, without actually calling any LLM, and is used for debugging and testing. The main change in the latest version 0.5 is that simulated token usage statistics now include tool calls, the system prompt, and reasoning tokens. Users can run the plugin via uvx, combined with the -o thinking 1 --usage parameters, to view output that includes usage information.
AI overview of reporting · Updated 3 h前
Reporting timeline
Explore different aspects of the story through its reporting.
- Simon Willisonllm-echo 0.5
Simon Willison updated llm-echo, a debugging plugin for LLM that simulates a "model" that simply echoes back the input prompt without calling any LLM. The new version now includes tool calls, system prompts, and reasoning tokens in its simulated token usage counts, and can be run via uvx with -o thinking 1 --usage to view the output.
Story heat trend
Current heat9·Comparable peak10（10/19 09:00）·Comparable change over 24 hours –
The trend compares the same continuously observed sources, which may cover fewer sources than the current heat score. Hover or select the chart to inspect each hour; use the left and right arrow keys to navigate.