What it does
This page is Google Cloud's overview of its AI and machine learning products, grouped into generative AI, machine learning and MLOps, speech, text and language APIs, image and video APIs, document and data APIs, conversational AI, AI infrastructure and a consulting programme. On the generative side it lists Gemini for your business, for discovering, creating, sharing and running agents, Agent Studio for prototyping and testing prompts and models, vector embeddings and search across BigQuery, AlloyDB and Spanner, and Agent Search for building search experiences. For machine learning it describes the Gemini platform for creating, training, tuning and deploying models, with over 200 models in its Model Garden, plus notebooks through Colab Enterprise or Workbench and BigQuery AI. The API products include Natural Language AI, Speech-to-Text, Text-to-Speech, Translation AI, Vision AI, Video AI and Document AI. CX Agent Studio covers conversational agents, and the infrastructure section covers TPUs, GPUs and CPUs, Google Kubernetes Engine and Cloud Run with on-demand GPUs. The page does not give prices. It notes that Cloud AI products follow Google Cloud's SLA policies and may have different latency or availability guarantees from other Google Cloud services.
Capabilities
- Agent Studio for prototyping and testing prompts and models
- Model Garden with 200+ proprietary, open and third-party models
- Speech-to-Text, Text-to-Speech, Translation, Natural Language, Vision and Video APIs
- Document AI for turning documents into structured data
- Vector search across BigQuery, AlloyDB and Spanner
- TPUs, GPUs, GKE and Cloud Run for training and serving
When to use it
- You want to see which Google Cloud AI service fits a task such as speech, vision, translation or document processing
- You need a managed platform to train, tune and deploy ML and generative AI models
- You are building agents or conversational experiences on Google Cloud
- You need AI-optimised compute for training or serving models
