Learn how to generate clean structured data like JSON with Claude using assistant message prefilling and stop sequences. Get raw data without explanatory text or markdown formatting.
Learn how semantic search uses text embeddings to find the most relevant chunks in a RAG pipeline. Understand embeddings, implement VoyageAI for generating embeddings, and discover how to match user queries with the right content.
Learn how to build the core evaluation pipeline for prompt testing. This guide covers the three essential functions—run_prompt, run_test_case, and run_eval—that process test cases through Claude and collect structured results for analysis.
Learn how to use fine-grained tool calling in Claude for faster, more granular streaming. Understand JSON validation, buffering delays, and when to enable fine-grained mode.
Learn how to use hooks in Claude Code to run commands before or after tool execution. Automate code formatting, testing, access control, and more with PreToolUse and PostToolUse hooks.