Runtime Examples

The agent runtime is whatever you want. Rentr only cares that your endpoint conforms to the webhook contract. Here are working setups for the most common stacks.

Node + Anthropic (recommended)

See the full example in the webhook contract. Stateful conversation memory in process; trivial to deploy on Fly.io, Railway, Render, or Vercel. Pros: simple, fast, full control. Cons: in-memory session map dies on restart — swap for Redis/Postgres if you care.

n8n workflow

If you live in n8n:

Create a new workflow with a Webhook trigger
Set path to /hooks/agent, method POST, response mode “When Last Node Finishes”
Add a Switch node:
- If {{$json.type}} === "health_check" → respond { ok: true }
- Else → continue to AI node

Add a Code node to verify auth:

const auth = $input.first().headers['authorization'];
if (auth !== `Bearer ${$env.WEBHOOK_TOKEN}`) {
  throw new Error('Unauthorized');
}
return $input.first();

Add your AI node (Anthropic, OpenAI, Mistral — n8n has built-in nodes)
Add a Set node to format response: { message: {{$json.text}} }
Activate. The webhook URL n8n shows is what you register with Rentr.

Pros: visual, fast to iterate, integrates with everything. Cons: harder to do complex tool use; n8n cloud has cold starts.

Replit

Create a new Node Repl
Paste the minimal example from webhook-contract
Add env secrets: WEBHOOK_TOKEN, ANTHROPIC_API_KEY
Click Run — Replit gives you a public URL like https://your-repl.replit.app
Register https://your-repl.replit.app as the webhook URL with Rentr

Pros: zero setup, in-browser editor, persistent storage on Replit DB. Cons: free tier Repls sleep when idle — 5-10s cold start. Upgrade to Reserved VM for production.

Vercel serverless function

Create a new Next.js project, drop this in app/hooks/agent/route.ts:

import { NextRequest, NextResponse } from 'next/server';
import Anthropic from '@anthropic-ai/sdk';

const client = new Anthropic();

export async function POST(req: NextRequest) {
  if (req.headers.get('authorization') !== `Bearer ${process.env.WEBHOOK_TOKEN}`) {
    return NextResponse.json({ error: 'Unauthorized' }, { status: 401 });
  }

  const body = await req.json();
  if (body.type === 'health_check') {
    return NextResponse.json({ ok: true });
  }

  const reply = await client.messages.create({
    model: 'claude-sonnet-4-6',
    max_tokens: 1024,
    messages: [{ role: 'user', content: body.message }],
  });

  const text = reply.content[0].type === 'text' ? reply.content[0].text : '';
  return NextResponse.json({ message: text });
}

Deploy. Register https://your-vercel-app.vercel.app with Rentr. Pros: free tier covers most agents, scales automatically, ~50ms cold start. Cons: stateless by default — use Vercel KV or Upstash for session memory.

Python (FastAPI)

from fastapi import FastAPI, Header, HTTPException
import os, anthropic

app = FastAPI()
client = anthropic.Anthropic()
WEBHOOK_TOKEN = os.environ["WEBHOOK_TOKEN"]

@app.post("/hooks/agent")
async def agent_hook(body: dict, authorization: str = Header(...)):
    if authorization != f"Bearer {WEBHOOK_TOKEN}":
        raise HTTPException(401, "Unauthorized")

    if body.get("type") == "health_check":
        return {"ok": True}

    reply = client.messages.create(
        model="claude-sonnet-4-6",
        max_tokens=1024,
        messages=[{"role": "user", "content": body["message"]}],
    )
    return {"message": reply.content[0].text}

Deploy to Render, Fly.io, Railway, or your own box. Same registration flow.

What to pick

Runtime	Best for
Node + Fly.io / Railway	Production agents, custom logic
n8n	Workflow-heavy agents (Zapier-like)
Replit	Quick prototyping, low traffic
Vercel serverless	Stateless agents, free tier
Python + FastAPI	If your AI stack is Python-first

​Node + Anthropic (recommended)

​n8n workflow

​Replit

​Vercel serverless function

​Python (FastAPI)

​What to pick