The rig-gemini-grpc crate provides the following capabilities:

• Full completion support with streaming
• Embedding generation
• Tool calling support
• Reasoning and thought signatures
• Image input support

Rig is architected to separate provider/backend contracts from agent orchestration through two main components:

• rig-core: Contains provider-neutral messages, completion models, portable tools, memory, and vector-store contracts, along with built-in provider mappings.
• rig-agent: Contains the agent orchestration logic, including the builder, prompt/streaming traits, typed hooks, contextual tools, extraction, and the serializable AgentRun state machine. This is enabled by default.

The root rig crate re-exports both, allowing most users to simply depend on rig.

To use AWS Bedrock as a model provider with Rig, add both rig-bedrock and rig-core to your Cargo.toml dependencies. You can use cargo add to fetch the latest versions automatically.

[dependencies]
rig-bedrock = "0.4.5"
rig-core = "0.36.0"

Or via CLI

cargo add rig-bedrock rig-core

To use the Rig-SQLite vector store, add rig-sqlite and rig-core to your Cargo.toml. You will also need serde (with derive features) and serde_json to handle metadata.

You can install the latest versions using the following commands:

cargo add rig-sqlite rig-core serde_json
cargo add serde --features derive

[dependencies]
rig-sqlite = "0.2.6"
rig-core = "0.37.0"
serde = { version = "1", features = ["derive"] }
serde_json = "1"

To use Milvus with Rig, add rig-milvus to your Rust project. Note that rig-core must also be added as a dependency for intended usage.

cargo add rig-milvus

To use LanceDB as a vector store with Rig, add rig-lancedb and rig-core to your Cargo.toml. You can use cargo add to fetch the latest versions automatically.

Pre-requisite: If running locally, you must have protoc (the Protobuf Compiler) installed on your system.

cargo add rig-lancedb rig-core

The rig-helixdb integration requires a specific queries/schema configuration to function. To deploy a minimum viable configuration using the provided example files:

1. Navigate to the configuration directory: cd helixdb-cfg
2. Deploy the configuration: helix push dev

This deploys the necessary queries and schema into your HelixDB instance.

The rig-agent crate provides Rig's classic agent runtime, including builders, serializable sans-I/O run state, blocking and streaming drivers, typed hooks, contextual tools, and extraction.

While most applications should use the root rig facade (where this runtime is enabled by default), you can import construction and prompting explicitly from rig-agent and rig-core.

use rig_agent::prelude::*;
use rig_core::{client::ProviderClient, providers::openai};

let client = openai::Client::from_env()?;
let agent = client.agent(openai::GPT_5_2).build();
let answer = agent.prompt("Explain ownership briefly.").await?;

Rig-Bedrock requires AWS credentials to be available in your environment before initializing the client. Ensure the following environment variables are set:

• AWS_DEFAULT_REGION
• AWS_SECRET_ACCESS_KEY
• AWS_ACCESS_KEY_ID

export AWS_DEFAULT_REGION=us-east-1
export AWS_SECRET_ACCESS_KEY=.......
export AWS_ACCESS_KEY_ID=......