Skip to main content
Eden AI automatically caches responses for repeated identical requests. When a request matches a previous one (same model and same input), the cached response is returned at no additional cost.
This page describes response caching, which returns an earlier result. For LLM prefix reuse that still generates a fresh response, see Prompt Caching.

​
How It Works

  1. You send a request with a specific model and input.
  2. Eden AI processes the request and stores the result.
  3. If the same model + input combination is sent again, the cached result is returned immediately.
Cached responses are faster (no round-trip to the provider) and free (no additional credit charge).

​
Best Suited For

Caching works well with deterministic features where the same input always produces the same output:
  • Embeddings — identical text produces identical vectors
  • Moderation — same text yields the same classification
  • OCR — same document produces the same extracted text
  • Named entity recognition — same text, same entities

​
Not Suited For

Caching is less useful for non-deterministic outputs:
  • LLM chat completions — responses vary by design (temperature, randomness)
  • Image generation — same prompt can produce different images

​
Activating Caching

Caching can be enabled or disabled per project from the Eden AI dashboard:
  1. Go to app.edenai.run
  2. Navigate to your project settings
  3. Toggle Response Caching on or off
Caching is enabled by default. Once active, it works automatically — no changes to your API calls are needed.