Module driver 

Calls a model. A Model pairs a Wire (what to send and how to read the reply) with a Transport (how the payload travels). Its stream yields any operation’s events as a Streamed, and its call is that stream, finished: one decoder, one fold and one response constructor for both. The _observed twins take the observation context a bus records under.

use rig_core::completion::CompletionRequest;
use rig_core::driver::Model;
use rig_core::providers::openai::{self, OpenAI};

let model = OpenAI::from_env()?.with_http(http).responses(openai::GPT_5_2);
let response = model.call(CompletionRequest::new("Hello")).await?;

Structs

DynModel

A model of one operation with its wire and transport erased. Clones share the model. Built with Model::erase or From<Model<W, T>>, so a consumer takes impl Into<DynModel<Op>> and accepts either.

Exchange

What the driver tells a transport about one send.

Local

A wire whose payload is the request and whose frames are the reply’s steps, for an operation whose events the runtime builds itself.

Model

An endpoint of one provider: a wire bound to a transport.

Opened

One reply a transport opened: its frames, and the facts the transport owns.

Opening

One reply, opened when first polled.

Enums

Step

One step of a local runtime’s reply.

Traits

Transport

Sends a wire’s payloads and delivers its replies’ frames.

Functions

warn_unmodeled

Logs an unmodeled payload’s kind and serialized size, never its content. Callers must supply a structural kind label without sensitive data.