Overview
Simon Willison released research-llm-apis, a repository containing scripts and captured outputs for testing LLM API interactions. This is preparatory work for redesigning the abstraction layer in his LLM Python library to support new vendor features like server-side tool execution.
Changes by Version
2026-04-04
- Added curl commands for testing raw JSON responses from Anthropic, OpenAI, Gemini and Mistral APIs
- Added captured outputs from streaming and non-streaming modes across different scenarios
- Added research scripts generated by Claude Code to help design new LLM abstraction layer