Crate cortexai_fastly 

Rust AI Agents - Fastly Compute

Run AI agents on Fastly Compute@Edge.

This crate provides a Fastly-native implementation using:

• cortexai-llm-client for request/response logic
• Fastly SDK for HTTP requests

Usage

use cortexai_fastly::{FastlyAgent, FastlyAgentConfig};
use fastly::{Request, Response};

#[fastly::main]
fn main(req: Request) -> Result<Response, fastly::Error> {
    let config = FastlyAgentConfig::new(
        "openai",
        std::env::var("OPENAI_API_KEY").unwrap(),
        "gpt-4o-mini",
    );

    let mut agent = FastlyAgent::new(config, "llm-backend");
    let response = agent.chat("Hello!")?;

    Ok(Response::from_body(response.content))
}

Structs

FastlyAgent

AI Agent for Fastly Compute

FastlyAgentConfig

Configuration for the Fastly agent

StreamingResponse

Streaming response iterator

Enums

FastlyAgentError

Errors that can occur in the Fastly agent

Functions

handle_chat_request

Create a simple chat completion handler for Fastly

Type Aliases

Result