Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.agipower.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Get generation

Encountering errors? See the API Error Codes Reference for a complete list of error types and troubleshooting steps.
GET https://api.agipower.ai/v1/management/generation?id=<generation_id>
The previous endpoint GET /v1/generation is deprecated. Please migrate to the new URL GET /v1/management/generation. The old URL will be removed in a future release.
The Get generation endpoint is used to retrieve generation details, such as usage and costs.
This endpoint supports retrieving generation details for all API protocols, including OpenAI Chat Completions, OpenAI Responses, Anthropic, and Vertex AI.
This endpoint only supports billing queries for Pay As You Go API keys. If you call this endpoint with a subscription-plan API key (prefixed with sk-ss-v1-), billing-related fields (such as usage, ratingResponses, etc.) will not be returned.To retrieve billing information, please use a Pay As You Go API key. See:

​
Metering and Billing Information

​
Metering (Token Usage)

Metering data (e.g., token usage in the nativeTokens field) is returned synchronously with the request in the protocol’s native format:
  • OpenAI Chat Completions protocol: returned in the response usage field
  • OpenAI Responses protocol: returned in the response usage field
  • Anthropic protocol: returned in the response usage field
  • Vertex AI protocol: returned in the response usageMetadata field

​
Billing (Billing & Costs)

Billing data (cost-related fields such as usage, ratingResponses, etc.) is not currently returned synchronously with the request. After the request completes, you must query it via this endpoint 3–5 minutes later.
We’re improving and upgrading our billing architecture to enable synchronous billing data in responses as soon as possible. Stay tuned!

​
Request params

​
Authorization Header Required

Header parameters: 
Authorization: Bearer <AGIPower_API_KEY>
  • Name: Authorization
  • Format: Bearer &lt;API_KEY&gt;
  • Description: Your AGIPower API key
    • Pay As You Go API key: supports querying full metering and billing information
    • Subscription API key (prefixed with sk-ss-v1-): supports metering only; billing information is not supported

​
generate_id string Required

Query parameters: The generation id returned by AGIPower API endpoints. You can obtain it from:

​
Returns

​
api string

API type. Values vary by protocol:
  • chat.completions - OpenAI Chat Completions protocol
  • responses - OpenAI Responses protocol
  • messages - Anthropic protocol
  • generateContent - Vertex AI protocol

​
generationId string

The current generation id.

​
model string

Model ID.

​
createAt string

The time when the server received the inference request.

​
generationTime integer

Total duration of this inference from first token to completion, in milliseconds.

​
latency integer

Time to first token, in milliseconds.

​
nativeTokens object

Usage information consumed by this inference, including:
  • completion_tokens integer - Tokens used for the completion
  • prompt_tokens integer - Tokens used for the prompt
  • total_tokens integer - Total tokens
  • completion_tokens_details object - Completion token details
    • reasoning_tokens integer - Tokens used for reasoning
  • prompt_tokens_details object - Prompt token details
    • cached_tokens integer - Cached tokens

​
streamed boolean

Whether the response is streamed.

​
finishReason string

The reason the model stopped generating.

​
usage number

Credits consumed by this inference.

​
ratingResponses object

Billing response details, including:
  • billAmount number - Billed amount
  • discountAmount number - Discount amount
  • originAmount number - Original amount
  • priceVersion string - Price version
  • ratingDetails array - Billing detail items, each containing:
    • billAmount number - Billed amount
    • discountAmount number - Discount amount
    • feeItemCode string - Fee item code (e.g., completion, prompt)
    • originAmount number - Original amount
    • rate number - Rate

​
requestRetryTimes integer

Number of request retries.

​
finalRetry boolean

Whether this is the final retry.

GET /v1/management/generation

curl https://api.agipower.ai/v1/management/generation?id=<generation_id> \
  -H "Authorization: Bearer $AGIPower_API_KEY"

API Response

{
  "api": "chat.completions",
  "generationId": "gen_01abc123def456",
  "model": "openai/gpt-4o",
  "createAt": "2026-03-26T06:00:00.000Z",
  "generationTime": 3200,
  "latency": 500,
  "nativeTokens": {
    "completion_tokens": 128,
    "prompt_tokens": 32,
    "total_tokens": 160,
    "completion_tokens_details": {
      "reasoning_tokens": 0
    },
    "prompt_tokens_details": {
      "cached_tokens": 0
    }
  },
  "streamed": true,
  "finishReason": "stop",
  "usage": 0.0052,
  "ratingResponses": {
    "billAmount": 0.0052,
    "discountAmount": 0,
    "originAmount": 0.0052,
    "priceVersion": "v1",
    "ratingDetails": [
      {
        "billAmount": 0.0016,
        "discountAmount": 0,
        "feeItemCode": "prompt",
        "originAmount": 0.0016,
        "rate": 0.00005
      },
      {
        "billAmount": 0.0036,
        "discountAmount": 0,
        "feeItemCode": "completion",
        "originAmount": 0.0036,
        "rate": 0.00003
      }
    ]
  },
  "requestRetryTimes": 0,
  "finalRetry": false
}