Chaser Documentation

Connect to Chaser sessions with Puppeteer. The API is the same as connecting to a local browser — only the endpoint changes.

Installation

$ npm i puppeteer-core

Use puppeteer-core, not puppeteer. The full puppeteer package downloads its own Chromium, which you don’t need — the browser is on Chaser’s side.

Basic connection

1 import puppeteer from "puppeteer-core";
2 
3 // 1. Create a session
4 const { cdp_url, id } = await fetch(
5   "https://api.chaser.sh/v1/sessions",
6   {
7     method: "POST",
8     headers: {
9       Authorization: `Bearer ${process.env.CHASER_KEY}`,
10       "Content-Type": "application/json",
11     },
12     body: JSON.stringify({ ttl_seconds: 1800 }),
13   }
14 ).then((r) => r.json());
15 
16 // 2. Connect
17 const browser = await puppeteer.connect({ browserWSEndpoint: cdp_url });
18 const [page] = await browser.pages();
19 
20 // 3. Automate
21 await page.goto("https://www.cloudflare.com");
22 console.log(await page.title());
23 
24 // 4. Screenshot via the API
25 const png = await fetch(
26   `https://api.chaser.sh/v1/sessions/${id}/screenshot`,
27   {
28     method: "POST",
29     headers: { Authorization: `Bearer ${process.env.CHASER_KEY}` },
30   }
31 ).then((r) => r.arrayBuffer());
32 writeFileSync("page.png", Buffer.from(png));
33 
34 // 5. Clean up
35 await fetch(`https://api.chaser.sh/v1/sessions/${id}`, { method: "DELETE" });

Notes

Use browserWSEndpoint, not browserURL (which is deprecated).
puppeteer.connect() returns immediately. The session must be in ready status first.
As with patchright: don’t create new contexts, don’t override the user-agent, and use the /screenshot endpoint instead of page.screenshot().
Puppeteer’s bundled Chromium has known automation artifacts. For production use, prefer patchright (Playwright-based) which patches these at the browser level.