Axerity

The Cache API allows you to store and retrieve AI responses, reducing costs on repeated requests and improving latency.

Cost Savings

Cache hits cost only 0.01 credits compared to standard model pricing. Writing to cache is free.

Encrypted Storage

All cached data is encrypted at rest using AES-256 encryption.

How It Works

Write: Store a response with a unique key
Read: Retrieve cached responses by key
Auto-expire: Responses expire after TTL (default 7 days)

Endpoints

Method	Endpoint	Description
GET	`/api/v1/cache/:key`	Retrieve cached item
PUT	`/api/v1/cache/:key`	Store item in cache
DELETE	`/api/v1/cache/:key`	Delete cached item
GET	`/api/v1/cache`	List all cached keys

Store (PUT)

PUT https://axerity.com/api/v1/cache/:key

Request Body

Prop

Type

Example

curl -X PUT https://axerity.com/api/v1/cache/greeting-response \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "value": {
      "role": "assistant",
      "content": "Hello! How can I help you today?"
    },
    "ttl": 86400,
    "tags": ["greetings", "common"]
  }'

Response

{
  "key": "greeting-response",
  "created_at": "2024-01-15T10:30:00Z",
  "expires_at": "2024-01-16T10:30:00Z",
  "size_bytes": 52
}

Retrieve (GET)

GET https://axerity.com/api/v1/cache/:key

Example

curl https://axerity.com/api/v1/cache/greeting-response \
  -H "Authorization: Bearer YOUR_API_KEY"

Response (Cache Hit)

{
  "key": "greeting-response",
  "value": {
    "role": "assistant",
    "content": "Hello! How can I help you today?"
  },
  "created_at": "2024-01-15T10:30:00Z",
  "expires_at": "2024-01-16T10:30:00Z",
  "hit": true
}

Response (Cache Miss)

{
  "key": "greeting-response",
  "hit": false
}

Delete (DELETE)

DELETE https://axerity.com/api/v1/cache/:key

Example

curl -X DELETE https://axerity.com/api/v1/cache/greeting-response \
  -H "Authorization: Bearer YOUR_API_KEY"

List Keys (GET)

GET https://axerity.com/api/v1/cache

Query Parameters

Prop

Type

Example

curl "https://axerity.com/api/v1/cache?tag=greetings&limit=10" \
  -H "Authorization: Bearer YOUR_API_KEY"

Response

{
  "keys": [
    {
      "key": "greeting-response",
      "created_at": "2024-01-15T10:30:00Z",
      "expires_at": "2024-01-16T10:30:00Z",
      "size_bytes": 52,
      "tags": ["greetings", "common"]
    }
  ],
  "total": 1,
  "storage_used_bytes": 52
}

Best Practices

Use Consistent Keys

Create deterministic keys based on your request parameters:

function getCacheKey(model, messages) {
  const hash = crypto
    .createHash("sha256")
    .update(JSON.stringify({ model, messages }))
    .digest("hex")
    .slice(0, 16);
  return `chat-${model}-${hash}`;
}

Cache Common Queries

Ideal candidates for caching:

FAQ responses
Greeting messages
Static information lookups
Repeated summarization tasks

Set Appropriate TTL

Use Case	Recommended TTL
Static content	30 days
FAQ responses	7 days
Dynamic summaries	1 day
Real-time data	Don't cache

Pricing

Operation	Cost
Cache hit	0.01 credits
Cache miss	0 credits
Write	Free
Delete	Free
Storage	$0.10/GB/month

Free Tier

All users get 100 MB of free cache storage.

Cache API

On this page