Introduction

Start building AI applications with the OpenRouter API

Welcome, AI engineer! These docs cover guides, examples, references, and more to help you build applications using the 100+ models available via OpenRouter with the Orpheus library.

hello.rs

use orpheus::prelude::*;

fn main() -> anyhow::Result<()> {
    let client = Orpheus::from_env()?;

    let response = client
        .chat("Say hello to our friend!")
        .model("openai/gpt-4o")
        .send()?;

    println!("Model says: {}", response.content()?);

    Ok(())
}

output

Model says: Hello there! It's great to meet you. How can I assist you today?

Objective

Orpheus aims to be three things:

• Ergonomic: Ease of use is a top priority. We keep the learning curve low and the amount of interfaces to remember to a minimum.

• Fully Featured: Stay up to date with the latest capabilities offered by providers. This includes tool calling, reasoning, multimodality, or something entirely new.

• Not a framework: Avoid opaque abstractions 4 staying as close as possible to the underlying API while providing a clearer, more comfortable interface.

Key Features

Orpheus also comes with out-of-the-box support for:

• ⚡ Async

• 🌊 Streaming

• 🖼️ Images, PDF, and Audio

• 🔄 Model Fallbacks

• 🔍 Web Search

• 🛠️ Tool Calling

• 🔌 MCP

• 📋 Structured Outputs

• ⚙️ Provider Selection

• 💾 Prompt Caching

• 🔧 Message Transforms