Multi-modal

llm-connector provides native support for multi-modal content (text + images + documents).

Image from URL

use llm_connector::{LlmClient, types::{ChatRequest, Message, MessageBlock}};

let client = LlmClient::openai("sk-...", "https://api.openai.com/v1")?;

let request = ChatRequest::new("gpt-4o")
    .add_message(Message::new(
        llm_connector::types::Role::User,
        vec![
            MessageBlock::text("What's in this image?"),
            MessageBlock::image_url("https://example.com/image.jpg"),
        ],
    ));

let response = client.chat(&request).await?;
println!("Response: {}", response.content);

Image from Base64

let base64_data = "iVBORw0KGgoAAAANSUhEUgAAAAUA...";

let request = ChatRequest::new("gpt-4o")
    .add_message(Message::new(
        llm_connector::types::Role::User,
        vec![
            MessageBlock::text("Analyze this image"),
            MessageBlock::image_base64("image/jpeg", base64_data),
        ],
    ));

Image from File

use llm_connector::types::MessageBlock;

let block = MessageBlock::image_file("path/to/image.jpg").await?;

Anthropic-Style Image (URL with detail)

let block = MessageBlock::image_url_with_detail(
    "https://example.com/image.jpg",
    "high"
);

Provider Support

Provider	Text	Images	Documents
OpenAI	✅	✅	❌
Anthropic	✅	✅	✅
Google Gemini	✅	✅	✅
Aliyun	✅	✅	❌
Zhipu	✅	✅	❌
Other	✅	❌	❌

Examples

cargo run --example multi_modal    # Basic image understanding
cargo run --example zhipu_vision   # Zhipu image analysis