Google Cloud Vision

Cloud Vision is Google Cloud's set of pretrained image analysis APIs: label and scene detection, OCR in dozens of languages including handwriting, face and landmark detection, logo recognition, explicit-content moderation, and visual product search for retail catalogs. It runs as REST or gRPC services with client libraries for the major languages, reading images from Cloud Storage or request payloads.

Google has migrated much of its newer vision work into the Vertex AI stack, but Cloud Vision remains the pragmatic way to get perception features tomorrow rather than train a custom model next quarter. Pricing is per 1,000 units per feature, and the first 1,000 units each month are free for most features.

It suits teams that need dependable general-purpose detection without model operations. There is little control over the underlying models, custom training pushes you toward Vertex AI, and high-volume or face-related workloads carry cost and biometric compliance considerations.