Run Intelligence | Resemble

POST https://app.resemble.ai/api/v2/intelligence

Analyzes a media file and returns comprehensive intelligence. The API auto-detects whether the input is audio, video, or image.

By default the request is synchronous. If you provide a callback_url, the request becomes asynchronous automatically: the API returns 202 Accepted with the newly created intelligence record, then sends the final result to your callback URL with an outbound HTTP POST.

Request Body

Field	Type	Required	Description
`media_token`	string	No*	Token for uploaded media file (from secure upload).
`url`	string	No*	HTTPS URL to media file.
`detect_id`	string	No	UUID of an existing detect to associate.
`media_type`	string	No	Explicit type: `audio`, `image`, or `video` (auto-detected if not provided).
`callback_url`	string	No	URL that receives the final intelligence payload via HTTP `POST`. When present, the request runs asynchronously and returns `202 Accepted`.

*One of media_token or url is required.

Example

$ curl --request POST 'https://app.resemble.ai/api/v2/intelligence' \
>   -H 'Authorization: Bearer YOUR_API_TOKEN' \
>   -H 'Content-Type: application/json' \
>   --data '{
>     "url": "https://example.com/media/sample.wav"
>   }'

Asynchronous Example (`callback_url`)

$ curl --request POST 'https://app.resemble.ai/api/v2/intelligence' \
>   -H 'Authorization: Bearer YOUR_API_TOKEN' \
>   -H 'Content-Type: application/json' \
>   --data '{
>     "url": "https://example.com/media/sample.wav",
>     "callback_url": "https://example.com/webhooks/intelligence"
>   }'

Response

Asynchronous Response (`callback_url` provided)

When callback_url is present, the API accepts the job immediately and returns a pending intelligence record:

1 {
2   "success": true,
3   "item": {
4     "uuid": "abc123...",
5     "media_type": "audio",
6     "created_at": "2024-01-15T10:30:00Z"
7   }
8 }

The final result is then delivered to your callback URL via HTTP POST.

Callback Payload

On success, the callback payload matches the normal intelligence response shape:

1 {
2   "success": true,
3   "item": {
4     "uuid": "abc123...",
5     "media_type": "audio",
6     "description": {
7       "speaker_info": "Adult male speaker",
8       "language": "en",
9       "dialect": "en-US",
10       "emotion": "confident",
11       "speaking_style": "conversational",
12       "context": "Product marketing pitch",
13       "message": "Introducing a new feature",
14       "abnormalities": "No anomalies detected",
15       "transcription": "Welcome back to the Resemble product update...",
16       "translation": null,
17       "misinformation": "No misinformation detected",
18       "fraud": {
19         "type": "none",
20         "confidence": 95,
21         "reasoning": "No fraud indicators detected"
22       },
23       "liveness": {
24         "assessment": "real_person",
25         "confidence": 88,
26         "indicators": "Natural breathing patterns, consistent background noise"
27       },
28       "digitally_altered": {
29         "detected": false,
30         "confidence": 92,
31         "alterations": "No digital alterations detected"
32       }
33     },
34     "created_at": "2024-01-15T10:30:00Z"
35   }
36 }

On failure, the callback includes the persisted intelligence record plus an error field:

1 {
2   "success": false,
3   "item": {
4     "uuid": "abc123...",
5     "media_type": "audio",
6     "description": "Error generating Intelligence report.",
7     "created_at": "2024-01-15T10:30:00Z"
8   },
9   "error": "Gemini API request failed"
10 }

Audio/Video Response

1 {
2   "success": true,
3   "item": {
4     "uuid": "abc123...",
5     "media_type": "audio",
6     "description": {
7       "speaker_info": "Adult male speaker",
8       "language": "en",
9       "dialect": "en-US",
10       "emotion": "confident",
11       "speaking_style": "conversational",
12       "context": "Product marketing pitch",
13       "message": "Introducing a new feature",
14       "abnormalities": "No anomalies detected",
15       "transcription": "Welcome back to the Resemble product update...",
16       "translation": null,
17       "misinformation": "No misinformation detected",
18       "fraud": {
19         "type": "none",
20         "confidence": 95,
21         "reasoning": "No fraud indicators detected"
22       },
23       "liveness": {
24         "assessment": "real_person",
25         "confidence": 88,
26         "indicators": "Natural breathing patterns, consistent background noise"
27       },
28       "digitally_altered": {
29         "detected": false,
30         "confidence": 92,
31         "alterations": "No digital alterations detected"
32       }
33     },
34     "created_at": "2024-01-15T10:30:00Z"
35   }
36 }

Notes

One of media_token or url is required.
Requests without callback_url are processed synchronously and return 200 OK.
Requests with callback_url are processed asynchronously and return 202 Accepted.
Use Get Intelligence to poll for the final result if you do not want to rely on callbacks.

Image Response

1 {
2   "success": true,
3   "item": {
4     "uuid": "def456...",
5     "media_type": "image",
6     "description": {
7       "scene_description": "Detailed description of what the image depicts",
8       "subjects": "People, objects, locations identified in the image",
9       "authenticity_analysis": "Visual artifacts, manipulation indicators, AI generation signs",
10       "context_and_setting": "Environmental, temporal, situational context",
11       "abnormalities": "Visual anomalies indicating AI generation or manipulation",
12       "misinformation": "Misinformation and disinformation analysis results",
13       "fraud": {
14         "type": "none",
15         "confidence": 90,
16         "reasoning": "No fraud indicators detected"
17       },
18       "liveness": {
19         "assessment": "real_person",
20         "confidence": 85,
21         "indicators": "Natural skin texture, consistent 3D geometry"
22       },
23       "digitally_altered": {
24         "detected": false,
25         "confidence": 88,
26         "alterations": "No digital alterations detected"
27       }
28     },
29     "created_at": "2024-01-15T10:30:00Z"
30   }
31 }

Response Fields

Audio/Video Description Fields

Field	Type	Description
`speaker_info`	string	Information about the speaker(s)
`language`	string	Detected language code
`dialect`	string	Detected dialect
`emotion`	string	Detected emotional tone
`speaking_style`	string	Speaking style classification
`context`	string	Contextual analysis of the content
`message`	string	Summary of the message conveyed
`abnormalities`	string	AI generation or manipulation anomalies
`transcription`	string	Full transcription of audio content
`translation`	string \| null	Translation if non-English content detected
`misinformation`	string	Misinformation and disinformation analysis
`fraud`	object	Fraud analysis result
`fraud.type`	string	Fraud classification: `none`, `vishing`, `impersonation`, `romance_scam`, `tech_support_scam`, `financial_fraud`, `extortion`, `political_manipulation`, `synthetic_media_fraud`, `insurance_fraud`, `employment_fraud`, `other`
`fraud.confidence`	number	Confidence score (0-100)
`fraud.reasoning`	string	Explanation of the fraud assessment
`liveness`	object	Liveness detection result
`liveness.assessment`	string	One of `real_person`, `not_real_person`, `inconclusive`
`liveness.confidence`	number	Confidence score (0-100)
`liveness.indicators`	string	Biometric liveness indicators found
`digitally_altered`	object	Digital alteration detection result
`digitally_altered.detected`	boolean	Whether manual digital editing was detected
`digitally_altered.confidence`	number	Confidence score (0-100)
`digitally_altered.alterations`	string	Description of alterations found

Image Description Fields

Field	Type	Description
`scene_description`	string	Detailed description of the image content
`subjects`	string	People, objects, and locations identified
`authenticity_analysis`	string	Visual artifacts and manipulation indicators
`context_and_setting`	string	Environmental and situational context
`abnormalities`	string	Visual anomalies indicating AI generation or manipulation
`misinformation`	string	Misinformation and disinformation analysis
`fraud`	object	Fraud analysis result
`fraud.type`	string	Fraud classification: `none`, `impersonation`, `synthetic_media_fraud`, `political_manipulation`, `insurance_fraud`, `other`
`fraud.confidence`	number	Confidence score (0-100)
`fraud.reasoning`	string	Explanation of the fraud assessment
`liveness`	object	Liveness detection result
`liveness.assessment`	string	One of `real_person`, `not_real_person`, `inconclusive`
`liveness.confidence`	number	Confidence score (0-100)
`liveness.indicators`	string	Biometric liveness indicators found
`digitally_altered`	object	Digital alteration detection result
`digitally_altered.detected`	boolean	Whether manual digital editing was detected
`digitally_altered.confidence`	number	Confidence score (0-100)
`digitally_altered.alterations`	string	Description of alterations found