Technique

Context Caching

Provider-side caching of long shared contexts (documents, codebases) across calls.

Definition

Context caching extends prompt caching to whole document or codebase contexts uploaded once and referenced cheaply across many queries. Gemini, Anthropic and OpenAI offer this as a way to amortise the cost of giant input prompts.

Common use cases

  • Codebase Q&A
  • Document chat
  • Long-context apps

Related terms

    Context Caching — AI Glossary | Railwail