@runbit/openai

An OpenAI-backed ChatModel for Runbit. Zero dependencies — uses global fetch.

npm install @runbit/openai @runbit/runtime

import { openai } from "@runbit/openai";
import { agent } from "@runbit/runtime";

const model = openai("gpt-4o-mini"); // reads OPENAI_API_KEY
const assistant = agent({ model, tools: [weather] });

Options

openai("gpt-4o-mini", {
  apiKey: process.env.OPENAI_API_KEY, // default
  baseURL: "https://api.openai.com/v1", // point at any OpenAI-compatible API
  temperature: 0.7,
  maxTokens: 1024,
});

Supports tool calling and reports token usage (so @runbit/devtools can show cost). Pass the result straight to agent({ model }).

Part of Runbit — the AI Runtime. Everything executes through one small, event-driven runtime; agents, workflows, RAG, and observability are packages on top.

• Umbrella: @runbit/runtime
• License: MIT