Rust SDK

The Trunkate AI Rust SDK offers the absolute lowest-latency, zero-cost abstraction bindings targeting the reqwest and tokio asynchronous runtimes.

Installation


cargo add trunkate-ai

Usage (Universal Budget)

Trunkate’s budget parameter uses a StringOrInt enum to support both absolute tokens and proactive percentage buffers.


use trunkate_ai::{Client, StringOrInt};
 
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let client = Client::new(None); 
    let text = "Your very long context...";
    
    // OPTION A: Fixed Token Budget (Integer)
    let res = client.optimize(text, None, Some(StringOrInt::Int(500)), None).await?;
    
    // OPTION B: Proactive Percentage Buffer (String)
    let res2 = client.optimize(text, None, Some(StringOrInt::String("20%".to_string())), None).await?;
    
    println!("Optimized: {}", res.text);
    Ok(())
}

Parameter Reference

Parameter	Type	Required	Description
`text`	`String`	Yes	The natural language prompt to optimize.
`task`	`Option<String>`	No	Optional guidance for preservation ranking.
`budget`	`Option<StringOrInt>`	No	Absolute tokens (`Int(500)`) or percentage (`String("20%")`).
`model`	`Option<String>`	No	Target LLM (default: `gpt-4o`).