If we don’t have the data to train our ML, we can use Google’s pre-trained APIs. These allow you not to worry about the implementation, but provided a standardized way to interact with a service. They save time and effort.

Some APIs provided by GCP are:

• Natural Language API: recognize entities and sentiment in a sentence
• Speech-to-Text API
• Text-to-Speech API
• Translation API
• Vision API: recognize content in images
• Vision intelligence API: motion and action in video
• Document AI API: text extraction and form parser
• Document warehouse API
• Dialogflow API: builds conversational interfaces
• Generative APIs: Gemini multimodal, embeddings multimodal, Gemini for text and chat, Imagen for image, Chirp for speech, Codey for code