Search across all documentation pages
10 pages in this section.
Understand how Claude's prompt caching works. Learn why byte-for-byte prefix matching and breakpoint placement are crucial for cost and latency savings.
Learn prompt caching basics with 8 Python examples. Discover how to add cache_control breakpoints, check cache creation, and understand ephemeral caching.
Learn to place cache_control breakpoints effectively in Claude API requests. Understand the order of tools, system prompts, and messages for optimal caching.
Compare Claude's 5-minute and 1-hour cache TTL options, their write costs, and how to select them for optimal prompt caching.
Verify cache hits by inspecting cache_read_input_tokens and cache_creation_input_tokens in the response usage object. Learn to implement a lightweight check.
Learn best practices for Claude prompt caching, covering breakpoint placement, choosing a TTL, and avoiding silent misses that erase savings.
A single-page roundup of every highlight bullet from the 9 pages in the Prompt Caching section, grouped by source page so you can scan all 53 takeaways without opening each article individually.