ReqRun with Next.js Route Handlers

Runnable Next.js App Router example that submits OpenAI-compatible work through ReqRun instead of calling the model provider directly.

What this repo shows

• a browser form talking to local Next.js route handlers
• a route handler that submits work to POST /v1/chat/completions
• a second route handler that proxies GET /v1/requests/{id}
• idempotency keys and async request ids without exposing the ReqRun API key to the browser

Why this is client-only

This repo does not contain ReqRun backend code. It is only a thin app-side integration that talks to the hosted ReqRun API at https://api.reqrun.com.

Prerequisites

• Node.js 20+
• a ReqRun project API key from https://app.reqrun.com

Setup

1. Copy the env template:

cp .env.example .env.local

2. Fill in:

REQRUN_API_KEY=REQRUN_LIVE_YOUR_PROJECT_KEY_HERE
REQRUN_BASE_URL=https://api.reqrun.com

3. Install dependencies:

npm install

4. Run the app:

npm run dev

5. Open http://localhost:3000

Run flow

The page sends your prompt to /api/run, which then sends this payload to ReqRun:

{
  "model": "gpt-5-nano",
  "messages": [
    { "role": "system", "content": "Return concise, useful output." },
    { "role": "user", "content": "Your prompt here" }
  ],
  "wait": false,
  "idempotency_key": "nextjs-demo-job-001"
}

When ReqRun returns an async response, the page can then call /api/requests/{id} to fetch the durable status by request id.

ReqRun docs

• Product docs: https://www.reqrun.com/docs
• Quickstart: https://www.reqrun.com/docs/quickstart
• API reference: https://www.reqrun.com/docs/api