vocabulary: "1.0.0" info: provider: Amazon Rekognition description: Unified vocabulary mapping operational API dimensions and capability workflows for Amazon Rekognition computer vision service. created: "2026-04-19" modified: "2026-04-19" operational: apis: - namespace: rekognition version: "2016-06-27" baseUrl: https://rekognition.amazonaws.com status: active description: Amazon Rekognition image and video analysis API resources: - name: labels api: rekognition description: Object, scene, and concept labels detected in images and videos actions: - detect - start-detection - get-detection - name: faces api: rekognition description: Faces detected and analyzed in images and video actions: - detect - compare - index - search - search-by-image - associate - disassociate - delete - list - name: collections api: rekognition description: Searchable face collections for identity management actions: - create - list - describe - delete - name: users api: rekognition description: User profiles within face collections actions: - create - delete - list - search - search-by-image - name: text api: rekognition description: Text detected in images and videos actions: - detect - start-detection - get-detection - name: moderation-labels api: rekognition description: Inappropriate or unsafe content labels actions: - detect - start-detection - get-detection - name: celebrities api: rekognition description: Celebrities recognized in images and videos actions: - recognize - get-info - start-recognition - get-recognition - name: liveness-sessions api: rekognition description: Face liveness verification sessions actions: - create - get-results - name: stream-processors api: rekognition description: Streaming video analysis processors actions: - create - start - stop - describe - list - delete - update - name: projects api: rekognition description: Custom Labels machine learning projects actions: - create - delete - describe - name: media-analysis-jobs api: rekognition description: Batch media analysis jobs actions: - start - get - list - name: custom-labels api: rekognition description: Custom trained labels detected in images actions: - detect actions: - name: detect httpMethods: - POST pattern: read description: Synchronously analyze an image and return detected entities - name: start httpMethods: - POST pattern: write description: Start an asynchronous video analysis job - name: get httpMethods: - POST pattern: read description: Retrieve results of an asynchronous job - name: create httpMethods: - POST pattern: write description: Create a new resource - name: delete httpMethods: - POST pattern: destructive description: Delete an existing resource - name: list httpMethods: - POST pattern: read description: List resources with optional pagination - name: describe httpMethods: - POST pattern: read description: Get detailed information about a specific resource - name: search httpMethods: - POST pattern: read description: Search for matching entities in a collection - name: index httpMethods: - POST pattern: write description: Add entities to a searchable collection - name: compare httpMethods: - POST pattern: read description: Compare two entities for similarity - name: recognize httpMethods: - POST pattern: read description: Identify known entities in media schemas: core: - name: Image description: Input image as S3 object or base64 bytes keyProperties: - Bytes - S3Object - name: S3Object description: S3 bucket and key reference for stored media keyProperties: - Bucket - Name - Version - name: BoundingBox description: Bounding box coordinates for detected entities keyProperties: - Width - Height - Left - Top - name: Label description: Detected label with name and confidence keyProperties: - Name - Confidence - Instances - Parents - name: FaceDetail description: Facial attributes detected on a face keyProperties: - BoundingBox - AgeRange - Emotions - Gender - Confidence video: - name: Video description: Video stored in S3 for analysis keyProperties: - S3Object - name: NotificationChannel description: SNS notification for async job completion keyProperties: - SNSTopicArn - RoleArn parameters: pagination: - name: NextToken description: Token for the next page of results - name: MaxResults description: Maximum number of results to return identifiers: - name: CollectionId description: Face collection identifier - name: FaceId description: UUID of an indexed face - name: UserId description: User profile identifier - name: JobId description: Async video analysis job identifier - name: SessionId description: Face liveness session identifier - name: ProjectVersionArn description: ARN of a Custom Labels model version filters: - name: MinConfidence description: Minimum confidence threshold for results - name: FaceMatchThreshold description: Minimum similarity threshold for face matches - name: SimilarityThreshold description: Minimum similarity score for face comparison enums: jobStatus: - IN_PROGRESS - SUCCEEDED - FAILED orientationCorrection: - ROTATE_0 - ROTATE_90 - ROTATE_180 - ROTATE_270 emotionTypes: - HAPPY - SAD - ANGRY - CONFUSED - DISGUSTED - SURPRISED - CALM - UNKNOWN - FEAR textTypes: - LINE - WORD qualityFilter: - NONE - AUTO - LOW - MEDIUM - HIGH livenessStatus: - CREATED - IN_PROGRESS - SUCCEEDED - FAILED - EXPIRED authentication: schemes: - name: AWS Signature Version 4 type: apiKey header: Authorization description: AWS SigV4 authentication using access key and secret capability: workflows: - name: Computer Vision Workflows file: capabilities/computer-vision-workflows.yaml description: Unified computer vision, identity verification, and content moderation workflows apisCombined: - rekognition toolCount: 15 personas: - Application Developer - Security Engineer - Content Moderator personas: - id: application-developer name: Application Developer description: Developers building apps with computer vision features such as image search, face login, and text extraction. workflows: - Computer Vision Workflows - id: security-engineer name: Security Engineer description: Security teams using facial recognition and liveness detection for identity verification and fraud prevention. workflows: - Computer Vision Workflows - id: content-moderator name: Content Moderator description: Trust and safety teams moderating user-generated content for inappropriate or explicit imagery. workflows: - Computer Vision Workflows domains: - name: Image Analysis description: Synchronous analysis of images for labels, faces, text, and moderation resources: - labels - faces - text - moderation-labels - celebrities - custom-labels - name: Identity Verification description: Face comparison, collection search, and liveness detection for user identity resources: - faces - collections - liveness-sessions - users - name: Video Analysis description: Asynchronous analysis of stored and streaming video resources: - labels - faces - text - moderation-labels - celebrities - stream-processors - media-analysis-jobs namespaces: - name: rekognition type: consumed baseUri: https://rekognition.amazonaws.com - name: computer-vision-api type: rest-exposed port: 8080 - name: computer-vision-mcp type: mcp-exposed port: 9080 binds: - name: AWS_ACCESS_KEY_ID description: AWS access key for authenticating Rekognition API calls workflows: - Computer Vision Workflows - name: AWS_SECRET_ACCESS_KEY description: AWS secret key for authenticating Rekognition API calls workflows: - Computer Vision Workflows - name: AWS_REGION description: AWS region for Rekognition API endpoint workflows: - Computer Vision Workflows crossReference: - resource: labels operations: - detectLabels - startLabelDetection - getLabelDetection workflows: - Computer Vision Workflows personas: - Application Developer - resource: faces operations: - detectFaces - compareFaces - indexFaces - searchFacesByImage workflows: - Computer Vision Workflows personas: - Application Developer - Security Engineer - resource: liveness-sessions operations: - createFaceLivenessSession - getFaceLivenessSessionResults workflows: - Computer Vision Workflows personas: - Security Engineer - resource: moderation-labels operations: - detectModerationLabels workflows: - Computer Vision Workflows personas: - Content Moderator