all agent api update benchmark dataset deployment fine tuning inference library model new model open source optimization prompt engineering quantization rag reinforcement learning research testing tool training tutorial visualization workflow
llm-echo 0.5a0
Simon Willison · 16h ago · 7 · tool testing open source

A new testing plugin provides a fake LLM model ('echo') that echoes prompts without actual inference, enabling developers to write automated tests for LLM-based applications. The tool supports faking reasoning blocks and JSON responses, streamlining test development workflows.