Website Screenshots in Laravel with the Http Facade and a Queued Job

Browsershot is the well-known way to take screenshots in Laravel, and it drags Node.js, Puppeteer and a Chrome binary into your PHP deployment. On shared hosting or a small container that is often the end of the story. The alternative is to treat capture as a remote call: one request with the Http facade, the rest is normal Laravel.

Configuration

// config/services.php
'snaprender' => [
    'key' => env('SNAPRENDER_API_KEY'),
    'endpoint' => 'https://app.snap-render.com/v1/screenshot',
],

# .env
SNAPRENDER_API_KEY=sk_live_your_key

A controller action

<?php

namespace App\Http\Controllers;

use Illuminate\Http\Request;
use Illuminate\Support\Facades\Http;

class ScreenshotController extends Controller
{
    public function show(Request $request)
    {
        $request->validate(['url' => ['required', 'url']]);

        $response = Http::withHeaders(['X-API-Key' => config('services.snaprender.key')])
            ->timeout(60)
            ->get(config('services.snaprender.endpoint'), [
                'url' => $request->string('url'),
                'format' => 'png',
                'full_page' => 'true',
                'block_cookie_banners' => 'true',
                'cache_ttl' => 86400,
            ]);

        if (! $response->successful()) {
            return response()->json(['error' => 'capture failed', 'detail' => $response->body()], 502);
        }

        return response($response->body(), 200)
            ->header('Content-Type', 'image/png')
            ->header('Cache-Control', 'public, max-age=86400');
    }
}

Route it and /screenshot?url=https://example.com streams a full-page PNG. timeout(60) keeps a slow page from tying up a PHP-FPM worker indefinitely.

A queued job with storage

Captures for bookmarks, monitoring or reports belong in the queue:

<?php

namespace App\Jobs;

use Illuminate\Bus\Queueable;
use Illuminate\Contracts\Queue\ShouldQueue;
use Illuminate\Foundation\Bus\Dispatchable;
use Illuminate\Queue\InteractsWithQueue;
use Illuminate\Queue\SerializesModels;
use Illuminate\Support\Facades\Http;
use Illuminate\Support\Facades\Storage;

class CapturePage implements ShouldQueue
{
    use Dispatchable, InteractsWithQueue, Queueable, SerializesModels;

    public int $tries = 3;
    public int $backoff = 30;

    public function __construct(public string $url, public string $name) {}

    public function handle(): void
    {
        $response = Http::withHeaders(['X-API-Key' => config('services.snaprender.key')])
            ->timeout(60)
            ->get(config('services.snaprender.endpoint'), [
                'url' => $this->url,
                'format' => 'jpeg',
                'full_page' => 'true',
                'cache_ttl' => 86400,
            ]);

        if ($response->status() === 429) {
            $this->release(60); // burst or quota exhausted: try again in a minute
            return;
        }

        $response->throw();

        Storage::disk('s3')->put("captures/{$this->name}.jpeg", $response->body());
    }
}

Dispatch with CapturePage::dispatch($url, $name). Swap s3 for any disk in config/filesystems.php.

For lists of URLs, the API's batch endpoint takes up to 50 per request and calls a webhook when the set is done; a single job that submits the batch and a controller that receives the webhook replaces a loop of jobs.

Caching on both sides

cache_ttl on the request makes repeated captures of the same URL and options free: they come from the API's cache and do not count against the monthly quota. On your side, cache the storage path:

$path = Cache::remember("shot:$url", 86400, fn () => $this->capture($url));

PDFs and other formats

format=pdf returns a page image inside a PDF, which is right for archives and wrong for text-searchable documents. To render your own Blade output as an image or PDF, POST the HTML to the same endpoint instead of a URL (see the HTML to PDF API).

Related: PHP quickstart, batch screenshots, webhooks.