ya
Yandex Vision is a computer vision service from Yandex Cloud providing OCR text extraction, image classification, face detection, and multi-feature batch image analysis.
With the Yandex Vision integration in AACFlow, you can:
- OCR (Recognize Text): Extract text from images and documents
- Classify Image: Categorize images using pre-trained classification models
- Detect Faces: Locate and analyze faces in images
- Batch Analyze (Advanced): Run multiple analysis features in a single request
This integration enables document digitization, content moderation, identity verification, and image data extraction workflows.
Usage Instructions
Integrate Yandex Vision into the workflow to analyze images. Requires a Yandex Cloud IAM token. Use the Yandex Cloud IAM block to obtain tokens programmatically. Images must be provided as base64-encoded data.
Tools
yandex_vision_ocr
Extract text from an image using OCR
Input
| Parameter | Type | Required | Description |
|---|---|---|---|
iamToken | string | Yes | Yandex Cloud IAM token |
imageData | string | Yes | Base64-encoded image data |
folderId | string | Yes | Yandex Cloud folder ID |
languageCodes | string | No | Language codes for OCR (e.g., ru,en) |
mimeType | string | No | Image MIME type (default: image/jpeg) |
Output
| Parameter | Type | Description |
|---|---|---|
text | string | Extracted text content |
blocks | array | Text blocks with position metadata |
yandex_vision_classify
Classify an image
Input
| Parameter | Type | Required | Description |
|---|---|---|---|
iamToken | string | Yes | Yandex Cloud IAM token |
imageData | string | Yes | Base64-encoded image data |
Output
| Parameter | Type | Description |
|---|---|---|
properties | array | Classification results with labels and probabilities |
yandex_vision_faces
Detect faces in an image
Input
| Parameter | Type | Required | Description |
|---|---|---|---|
iamToken | string | Yes | Yandex Cloud IAM token |
imageData | string | Yes | Base64-encoded image data |
Output
| Parameter | Type | Description |
|---|---|---|
faces | array | Detected faces with bounding box coordinates |
count | number | Number of faces detected |
yandex_vision_analyze
Batch analyze image with multiple features
Input
| Parameter | Type | Required | Description |
|---|---|---|---|
iamToken | string | Yes | Yandex Cloud IAM token |
imageData | string | Yes | Base64-encoded image data |
Output
| Parameter | Type | Description |
|---|---|---|
results | array | Analysis results for each requested feature |
Links
- Official API: https://yandex.cloud/ru/docs/vision/
- AACFlow block source: apps/aacflow/blocks/blocks/yandex_vision.ts