ReqRun
Sign inGet started

SDK

Familiar OpenAI shape, smaller reliability surface.

The SDK keeps chat.completions.create and adds one status method for async request ids.

Current status

The canonical public package is @reqrun/sdk and it is now available on npm. The hosted HTTP API and the runnable example repos remain the best lower-level references when you want to inspect the exact signed request shape.

Planned client shape

TypeScript
import { ReqRun } from "@reqrun/sdk";

const reqrun = new ReqRun({
  apiKey: process.env.REQRUN_API_KEY!,
  signingSecret: process.env.REQRUN_SIGNING_SECRET!,
  baseURL: "https://api.reqrun.com",
});

const response = await reqrun.chat.completions.create({
  model: "gpt-5-nano",
  messages: [{ role: "user", content: "Hello" }],
  wait: true,
});

const request = await reqrun.requests.get("rr_abc123");

Async fallback

Use wait=false for background work, or handle the async object when wait=true cannot finish before the timeout.

TypeScript
const response = await reqrun.chat.completions.create({
  model: "gpt-5-nano",
  messages: [{ role: "user", content: "Run this background LLM task." }],
  wait: false,
  idempotency_key: "task-123",
});

if (response.object === "chat.completion.async") {
  await saveRequestId(response.id);
}

Recommended public path today

  • Use the hosted API at https://api.reqrun.com
  • Use POST /v1/chat/completions for request submission
  • Use GET /v1/requests/{id} for async status lookup
  • Use a ReqRun project API key plus the signing secret shown when the key was created

API surface

  • new ReqRun({ apiKey, signingSecret, baseURL? })
  • reqrun.chat.completions.create(params)
  • reqrun.requests.get(id)