---
title: Context compression vs. prompt caching
url: https://compresr.ai/compare/prompt-caching
description: Prompt caching and context compression are complementary: caching reuses repeated prefixes; compression shrinks unique context.
---

# Prompt caching and context compression are complementary, not rivals: caching reuses repeated prefixes; compression shrinks unique context.

> Prompt caching discounts tokens you have already sent; context compression reduces the unique-per-request tokens you send at all. Use both together.

> Human-readable page: https://compresr.ai/compare/prompt-caching

## How they differ
- Prompt caching: reuses already-sent prefix tokens (still billed, at a discount). Best for stable, repeated context.
- Context compression: reduces the number of tokens sent, including unique-per-request context. Best for long, varying context.
- They compose: compress the unique context, cache the stable prefix.

## Related machine surfaces
- [/compare/llmlingua.md](/compare/llmlingua.md)
- [/glossary/context-compression.md](/glossary/context-compression.md)

## Provenance
Compresr Inc. is a Y Combinator W26 company built by four EPFL-trained founders in San Francisco, California and Europe (Switzerland).
Contact: [compresr.ai/contact](https://compresr.ai/contact).