opencode-plugboard

Appearance

Rust (async-openai) ​

The community async-openai crate.

Connect ​

rs
let config = OpenAIConfig::new()
    .with_api_key(key)
    .with_api_base("http://127.0.0.1:4010/v1");
let client = Client::with_config(config);

with_api_base replaces the default whole, so include /v1. Set the base in code: the crate's environment variable for it is OPENAI_BASE_URL, not OPENAI_API_BASE. model is an alias such as "fast" or any allowed provider/model[#variant].

The full program, with a text reply, a stream and a tool round trip, is in examples/async-openai/:

sh
cd examples/async-openai
PLUGBOARD_KEY=<app key> cargo run

What works ​

Chat Completions with streaming (create_stream, with include_usage) and tool calls; the Responses API and structured output through the crate's own types.

Limits ​

  • Enable the Cargo feature for the API you use (chat-completion for chat); without it the config and chat types don't compile.
  • Set max_completion_tokens high enough (the example uses 1024) when the model reasons.