Request
json
{
"model": "blink-1.0-flash",
"state": { "ticket": "Checkout is down for every customer." },
"questions": {
"department": {
"type": "choice",
"instructions": "Which team should handle `ticket`?",
"criteria": {
"billing": "Invoices and refunds",
"technical": "Software failures",
"sales": "New purchases"
}
},
"urgent": {
"type": "noul",
"instructions": "Is urgent attention needed?"
},
"severity": {
"type": "score",
"instructions": "Rate severity.",
"criteria": ["No impact", "Some customers affected", "All customers blocked"]
}
}
}| Field | Type | Notes |
|---|---|---|
model |
string | blink-1.0-flash (default) or blink-1.0-pro |
state |
any JSON | Required |
images |
string[] | Optional. One base64 PNG/JPEG/WebP data URL |
videos |
string[] | Optional. One base64 MP4 data URL — see Video inputs. Cannot be combined with images. |
audio |
string[] | Optional. One base64 WAV or MP3 data URL. Requires blink-1.0-pro. |
questions |
object | 1–16 questions |
Use the same blnkd API key for both models. Flash is the default. Pro adds audio support. The previous experimental IDs remain accepted as aliases; responses use the new model names. In the console, select the model below the question editor. Send one image, video, or audio attachment per request.
Pro supports 8,192 input tokens, including media and questions. It evaluates each question separately; usage includes the input tokens for each evaluation. Output tokens are zero because it returns classifier probabilities. Audio uses the first 30 seconds of the recording.
Response
Illustrative response — values vary with the input. Each answer uses the same ID as its question.
json
{
"model": "blink-1.0-flash",
"answers": {
"department": {
"type": "choice",
"choice": "technical",
"confidence": 0.91,
"probabilities": { "billing": 0.03, "technical": 0.91, "sales": 0.06 }
},
"urgent": { "type": "noul", "noul": 0.97 },
"severity": {
"type": "score",
"score": 1.86,
"confidence": 0.87,
"legend": { "0": "No impact", "1": "Some customers affected", "2": "All customers blocked" },
"probabilities": { "0": 0.01, "1": 0.12, "2": 0.87 }
}
},
"usage": { "input_tokens": 412, "output_tokens": 28 }
}