Crate adk_gemini 

adk-gemini

A Rust client library for Google’s Gemini 2.0 API.

Crate Organization

This crate is organized into domain-specific modules that align with the Gemini API’s capabilities:

• generation - Content generation, including text, images, and audio
• embedding - Text embedding generation for semantic analysis
• batch - Batch processing for multiple requests
• files - File upload and management
• cache - Content caching for reusable contexts
• safety - Content moderation and safety settings
• tools - Function calling and tool integration
• models - Core primitive types shared across modules
• prelude - Convenient re-exports of commonly used types

Quick Start

For most use cases, import from the prelude:

use adk_gemini::prelude::*;

For more specialized types, import them directly from the crate root or their respective modules.

Re-exports

pub use client::Error as ClientError;

pub use client::Gemini;

pub use client::GeminiBuilder;

pub use client::Model;

pub use generation::builder::ContentBuilder;

pub use generation::model::BlockReason;

pub use generation::model::Candidate;

pub use generation::model::CitationMetadata;

pub use generation::model::CitationSource;

pub use generation::model::FinishReason;

pub use generation::model::GenerateContentRequest;

pub use generation::model::GenerationConfig;

pub use generation::model::GenerationResponse;

pub use generation::model::GroundingChunk;

pub use generation::model::GroundingMetadata;

pub use generation::model::GroundingSegment;

pub use generation::model::GroundingSupport;

pub use generation::model::MultiSpeakerVoiceConfig;

pub use generation::model::PrebuiltVoiceConfig;

pub use generation::model::PromptFeedback;

pub use generation::model::PromptTokenDetails;

pub use generation::model::SpeakerVoiceConfig;

pub use generation::model::SpeechConfig;

pub use generation::model::ThinkingConfig;

pub use generation::model::ThinkingLevel;

pub use generation::model::UsageMetadata;

pub use generation::model::VoiceConfig;

pub use generation::model::WebGroundingChunk;

pub use embedding::builder::EmbedBuilder;

pub use embedding::model::BatchContentEmbeddingResponse;

pub use embedding::model::BatchEmbedContentsRequest;

pub use embedding::model::ContentEmbedding;

pub use embedding::model::ContentEmbeddingResponse;

pub use embedding::model::EmbedContentRequest;

pub use embedding::model::TaskType;

pub use safety::model::HarmBlockThreshold;

pub use safety::model::HarmCategory;

pub use safety::model::HarmProbability;

pub use safety::model::SafetyRating;

pub use safety::model::SafetySetting;

pub use tools::model::FunctionCall;

pub use tools::model::FunctionCallingConfig;

pub use tools::model::FunctionCallingMode;

pub use tools::model::FunctionDeclaration;

pub use tools::model::FunctionResponse;

pub use tools::model::FunctionResponsePart;

pub use tools::model::Tool;

pub use tools::model::ToolConfig;

pub use batch::Error as BatchError;

pub use batch::builder::BatchBuilder;

pub use batch::handle::BatchGenerationResponseItem;

pub use batch::handle::BatchHandle;

pub use batch::handle::BatchHandle as Batch;

pub use batch::handle::BatchStatus;

pub use batch::handle::Error as BatchHandleError;

pub use batch::model::BatchConfig;

pub use batch::model::BatchGenerateContentRequest;

pub use batch::model::BatchOperation;

pub use batch::model::BatchStats;

pub use batch::model::IndividualRequestError;

pub use batch::model::RequestMetadata;

pub use files::Error as FilesError;

pub use files::builder::FileBuilder;

pub use files::handle::FileHandle;

pub use files::model::File;

pub use files::model::FileState;

pub use cache::builder::CacheBuilder;

pub use cache::handle::CachedContentHandle;

pub use cache::model::CacheExpirationRequest;

pub use cache::model::CacheExpirationResponse;

pub use cache::model::CachedContent;

pub use cache::model::CreateCachedContentRequest;

pub use pricing::CostBreakdown;

pub use pricing::GeminiPricing;

pub use pricing::estimate_cost;

pub use pricing::estimate_cost_long;

pub use schema_adapter::GeminiSchemaAdapter;

Modules

backend

Backend abstraction for Gemini API providers (AI Studio REST vs Vertex AI). Backend abstraction for Gemini API providers.

batch

Batch processing for multiple generation requests Batch processing for multiple generation requests.

cache

Content caching for reusable contexts and system instructions Content caching for reusable contexts and system instructions.

client

Gemini API client, builder, model enum, and error types.

common

Common utilities and serialization helpers

embedding

Text embedding generation for semantic analysis

files

File upload and management File upload and management for the Gemini API.

generation

Content generation including text, images, and audio Content generation including text, images, and audio.

model_info

Model metadata discovery (list/get model info). Types for the Gemini Models API (models.list and models.get).

prelude

Convenient re-exports of commonly used types

pricing

Token cost estimation for Gemini models. Token pricing for Gemini models.

safety

Content moderation and safety settings Content moderation and safety settings for the Gemini API.

schema_adapter

Provider-specific schema normalization for Gemini function-calling API. Gemini-specific schema normalization adapter.

tools

Function calling and tool integration Function calling and tool integration for the Gemini API.

Structs

Blob

Core primitive types for building requests and parsing responses Blob for a message part

Content

Core primitive types for building requests and parsing responses Content of a message

FileDataRef

Core primitive types for building requests and parsing responses Reference to an external file by URI, used in Gemini wire format.

Message

Core primitive types for building requests and parsing responses Message in a conversation

Enums

Modality

Core primitive types for building requests and parsing responses Content modality type - specifies the format of model output

Part

Core primitive types for building requests and parsing responses Content part that can be included in a message

Role

Core primitive types for building requests and parsing responses Role of a message in a conversation