> ## Documentation Index
> Fetch the complete documentation index at: https://docs.keywordsai.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Create dataset log

Create an individual dataset log from unified format data. Datasets record end-to-end input/output of workflows, so both input and output can be arbitrary JSON structures.

## Authentication

* API key: `Authorization: Bearer <API key>`

## Parameters

<ParamField body="input" type="any" required>
  The input data - can be any JSON structure (objects, arrays, strings, etc.)
</ParamField>

<ParamField body="output" type="any" required>
  The output data - can be any JSON structure (objects, arrays, strings, etc.)
</ParamField>

<ParamField body="metadata" type="object">
  Additional metadata fields (model, log\_type, custom\_identifier, etc.)
</ParamField>

<ParamField body="metrics" type="object">
  Metric fields (prompt\_tokens, completion\_tokens, cost, latency, etc.)
</ParamField>

Note: Fields can be provided either at top-level or within metadata/metrics objects. This endpoint supports any workflow type - customer support, data processing, API transformations, calculations, etc.

## Request Example

<RequestExample>
  ```json Request theme={"system"}
  {
    "input": {
      "customer_id": "cust_12345",
      "query_type": "billing",
      "message": "I need help with my subscription",
      "priority": "high",
      "user_context": {
        "plan": "premium",
        "account_age_days": 365
      }
    },
    "output": {
      "response": "I'd be happy to help with your subscription. Let me look that up for you.",
      "agent_id": "agent_789",
      "resolution_time_seconds": 120,
      "actions_taken": ["account_lookup", "plan_verification"],
      "satisfaction_score": 4.5,
      "follow_up_required": true
    },
    "metadata": {
      "model": "gpt-4",
      "log_type": "chat",
      "session_id": "sess_abc123",
      "custom_identifier": "support-interaction-v2"
    },
    "metrics": {
      "prompt_tokens": 25,
      "completion_tokens": 18,
      "total_tokens": 43,
      "cost": 0.002,
      "latency": 1.8
    }
  }
  ```
</RequestExample>

## Examples

<CodeGroup>
  ```python Python theme={"system"}
  import requests

  url = "https://api.keywordsai.co/api/datasets/{dataset_id}/logs/"
  headers = {
      "Authorization": "Bearer YOUR_API_KEY",
      "Content-Type": "application/json"
  }

  data = {
      "input": {
          "customer_id": "cust_12345",
          "query_type": "billing",
          "message": "I need help with my subscription",
          "priority": "high",
          "user_context": {
              "plan": "premium",
              "account_age_days": 365
          }
      },
      "output": {
          "response": "I'd be happy to help with your subscription. Let me look that up for you.",
          "agent_id": "agent_789",
          "resolution_time_seconds": 120,
          "actions_taken": ["account_lookup", "plan_verification"],
          "satisfaction_score": 4.5,
          "follow_up_required": True
      },
      "metadata": {
          "model": "gpt-4",
          "log_type": "chat",
          "session_id": "sess_abc123",
          "custom_identifier": "support-interaction-v2"
      },
      "metrics": {
          "prompt_tokens": 25,
          "completion_tokens": 18,
          "total_tokens": 43,
          "cost": 0.002,
          "latency": 1.8
      }
  }

  response = requests.post(url, headers=headers, json=data)
  print(response.json())
  ```

  ```bash cURL theme={"system"}
  curl -X POST "https://api.keywordsai.co/api/datasets/{dataset_id}/logs/" \
    -H "Authorization: Bearer YOUR_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "input": {
        "customer_id": "cust_12345",
        "query_type": "billing",
        "message": "I need help with my subscription"
      },
      "output": {
        "response": "I'\''d be happy to help with your subscription.",
        "agent_id": "agent_789",
        "satisfaction_score": 4.5
      },
      "metadata": {
        "model": "gpt-4",
        "log_type": "chat"
      },
      "metrics": {
        "prompt_tokens": 25,
        "completion_tokens": 18,
        "cost": 0.002,
        "latency": 1.8
      }
    }'
  ```
</CodeGroup>

## Response

<ResponseExample>
  ```json 201 Created theme={"system"}
  {
    "message": "Dataset log request completed synchronously (debug mode)!",
    "log_data": {
      "unique_id": "2844b4f401854250aec6ee1ffd8f38be",
      "input": "{\"customer_id\": \"cust_12345\", \"query_type\": \"billing\", \"message\": \"I need help with my subscription\"}",
      "output": {
        "response": "I'd be happy to help with your subscription. Let me look that up for you.",
        "agent_id": "agent_789",
        "resolution_time_seconds": 120,
        "satisfaction_score": 4.5
      },
      "model": "gpt-4",
      "log_type": "chat",
      "prompt_tokens": 25,
      "completion_tokens": 18,
      "cost": 0.002,
      "latency": 1.8,
      "dataset_id": "6828c7e1-51fa-4b05-89be-382567fc2703",
      "status": "success",
      "session_id": "sess_abc123",
      "custom_identifier": "support-interaction-v2"
    }
  }
  ```
</ResponseExample>

## Errors

* 400 Bad Request — Invalid unified format data
* 401 Unauthorized — Missing/invalid authentication
* 404 Not Found — Dataset not found or not in your organization