REST and RPC reference for generative AI services

You can find the REST API and RPC reference pages for generative AI services in the Vertex AI reference documentation.

REST reference

Vertex AI has generative-AI-specific REST resources, as well as REST resources that are used in both generative AI and predictive AI use cases.

Generative AI REST resources

The following table lists the generative-AI-specific REST resources on Vertex AI.

Generative AI resource
Description
projects.locations.publishers.models.generateContent Generates content for Gemini.
projects.locations.publishers.models.streamGenerateContent Streams generated content for Gemini.
projects.locations.publishers.models.serverStreamingPredict Streams generated content for PaLM 2 and Codey.
projects.locations.publishers.models.predict Generates content for non-Gemini models.
projects.locations.publishers.models.countTokens (v1beta1) Gets token count and billable characters for a prompt.
projects.locations.tuningJobs Tunes Gemini models.
projects.locations.reasoningEngines (v1beta1) Deploy LangChain applications to Vertex AI.

Shared REST resources

The following table lists the shared REST resources on Vertex AI.

Shared REST resource
Description
projects.locations.pipelineJobs Used to tune non-Gemini models and for model evaluations.
projects.locations.batchPredictionJobs Used to send batch prompt request to supported models.
projects.locations.endpoints Used to deploy Model Garden models.

RPC reference

The following table lists the RPCs for generative AI.

RPC package
Description
cloud.ai.large_models.vision Large vision models.
cloud.ai.nl.llm.proto.service Large language models.
learning.genai.root Grounding and citations.