# Content Provenance Checks

## Create content provenance check

`content_provenance_checks.create(ContentProvenanceCheckCreateParams**kwargs)  -> ContentProvenanceCheck`

**post** `/content_provenance_checks`

Check whether an image or audio file contains known OpenAI provenance signals. [Learn more about content provenance](/api/docs/guides/content-provenance).

If `not_detected`, it means the tool did not find supported signals in the uploaded file. The content could still have been generated by OpenAI if the metadata was stripped or has evidence of tampering, the watermark was degraded, it comes from a legacy generation model, or it was created before provenance signals were available. Content could also still be AI-generated by another company's model, which the tool currently does not detect.

### Parameters

- `file: FileTypes`

  The image or audio file to check for supported OpenAI provenance signals.

### Returns

- `class ContentProvenanceCheck: …`

  - `created_at: int`

    The Unix timestamp, in seconds, when the provenance check was created.

  - `object: Literal["content_provenance_check"]`

    The object type. Always `content_provenance_check` for this endpoint.

    - `"content_provenance_check"`

  - `results: List[Result]`

    The provenance results that apply to the uploaded file. Image results include C2PA and SynthID; audio results include SynthID.

    - `class ResultC2PA: …`

      - `generated_at: Optional[str]`

        The UTC RFC 3339 timestamp recorded by the provenance signal for when the asset was generated, when available.

      - `issuer: Optional[str]`

        The C2PA manifest issuer, when available.

      - `model: Optional[str]`

        The OpenAI model recorded by the provenance signal, when available.

      - `outcome: Literal["detected", "not_detected"]`

        Whether a supported OpenAI C2PA provenance signal was detected.
        If `not_detected`, it means the tool did not find supported signals in the uploaded file. The content could still have been generated by OpenAI if the metadata was stripped or has evidence of tampering, the watermark was degraded, it comes from a legacy generation model, or it was created before provenance signals were available. Content could also still be AI-generated by another company's model, which the tool currently does not detect.

        - `"detected"`

        - `"not_detected"`

      - `type: Literal["c2pa"]`

        The provenance signal type. Always `c2pa`.

        - `"c2pa"`

      - `validation_state: Literal["trusted", "valid", "invalid", "not_present"]`

        The validation status of the C2PA manifest in the uploaded image.

        - `"trusted"`

        - `"valid"`

        - `"invalid"`

        - `"not_present"`

    - `class ResultSynthID: …`

      - `generated_at: Optional[str]`

        The UTC RFC 3339 timestamp recorded by the provenance signal for when the asset was generated, when available.

      - `model: Optional[str]`

        The OpenAI model recorded by the provenance signal, when available.

      - `outcome: Literal["detected", "not_detected"]`

        Whether a supported OpenAI SynthID watermark was detected.
        If `not_detected`, it means the tool did not find supported signals in the uploaded file. The content could still have been generated by OpenAI if the metadata was stripped or has evidence of tampering, the watermark was degraded, it comes from a legacy generation model, or it was created before provenance signals were available. Content could also still be AI-generated by another company's model, which the tool currently does not detect.

        - `"detected"`

        - `"not_detected"`

      - `type: Literal["synthid"]`

        The provenance signal type. Always `synthid`.

        - `"synthid"`

### Example

```python
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ.get("OPENAI_API_KEY"),  # This is the default and can be omitted
)
content_provenance_check = client.content_provenance_checks.create(
    file=b"Example data",
)
print(content_provenance_check.created_at)
```

#### Response

```json
{
  "created_at": 0,
  "object": "content_provenance_check",
  "results": [
    {
      "generated_at": "generated_at",
      "issuer": "issuer",
      "model": "model",
      "outcome": "detected",
      "type": "c2pa",
      "validation_state": "trusted"
    }
  ]
}
```

## Domain Types

### Content Provenance Check

- `class ContentProvenanceCheck: …`

  - `created_at: int`

    The Unix timestamp, in seconds, when the provenance check was created.

  - `object: Literal["content_provenance_check"]`

    The object type. Always `content_provenance_check` for this endpoint.

    - `"content_provenance_check"`

  - `results: List[Result]`

    The provenance results that apply to the uploaded file. Image results include C2PA and SynthID; audio results include SynthID.

    - `class ResultC2PA: …`

      - `generated_at: Optional[str]`

        The UTC RFC 3339 timestamp recorded by the provenance signal for when the asset was generated, when available.

      - `issuer: Optional[str]`

        The C2PA manifest issuer, when available.

      - `model: Optional[str]`

        The OpenAI model recorded by the provenance signal, when available.

      - `outcome: Literal["detected", "not_detected"]`

        Whether a supported OpenAI C2PA provenance signal was detected.
        If `not_detected`, it means the tool did not find supported signals in the uploaded file. The content could still have been generated by OpenAI if the metadata was stripped or has evidence of tampering, the watermark was degraded, it comes from a legacy generation model, or it was created before provenance signals were available. Content could also still be AI-generated by another company's model, which the tool currently does not detect.

        - `"detected"`

        - `"not_detected"`

      - `type: Literal["c2pa"]`

        The provenance signal type. Always `c2pa`.

        - `"c2pa"`

      - `validation_state: Literal["trusted", "valid", "invalid", "not_present"]`

        The validation status of the C2PA manifest in the uploaded image.

        - `"trusted"`

        - `"valid"`

        - `"invalid"`

        - `"not_present"`

    - `class ResultSynthID: …`

      - `generated_at: Optional[str]`

        The UTC RFC 3339 timestamp recorded by the provenance signal for when the asset was generated, when available.

      - `model: Optional[str]`

        The OpenAI model recorded by the provenance signal, when available.

      - `outcome: Literal["detected", "not_detected"]`

        Whether a supported OpenAI SynthID watermark was detected.
        If `not_detected`, it means the tool did not find supported signals in the uploaded file. The content could still have been generated by OpenAI if the metadata was stripped or has evidence of tampering, the watermark was degraded, it comes from a legacy generation model, or it was created before provenance signals were available. Content could also still be AI-generated by another company's model, which the tool currently does not detect.

        - `"detected"`

        - `"not_detected"`

      - `type: Literal["synthid"]`

        The provenance signal type. Always `synthid`.

        - `"synthid"`