AWS Lambda Chromium Binary Size Limit

Why your Puppeteer deployment fails and what to do about it

The Error

Unzipped size must be smaller than 262144000 bytes

# Or you might see:
Resource handler returned message: "Unzipped size must be smaller
than 262144000 bytes" (RequestToken: xxx, HandlerErrorCode: InvalidRequest)

This happens when your Lambda deployment package (including dependencies) exceeds 250 MB unzipped. Chromium alone is ~280 MB. You're stuck before you even start.

Why This Happens

Lambda Limits

250 MB unzipped deployment size
50 MB zipped upload limit
Lambda Layers count toward this

Chromium Size

Full Chromium: ~280 MB
chromium-min: ~45 MB (stripped)
+ Puppeteer/Playwright: ~50 MB

Even with stripped-down Chromium builds, you're constantly fighting the limit. Add your application code and any other dependencies, and you hit the wall fast.

Traditional "Solutions" (And Their Problems)

1. Use Lambda Layers

Split Chromium into a separate Layer to stay under the limit.

Problems:

Layers still count toward 250 MB total
Layer management adds deployment complexity
Version mismatches between Puppeteer and Chromium
Cold starts are still brutal (~8-15 seconds)

2. Use @sparticuz/chromium

A stripped-down Chromium build for Lambda (~45 MB).

npm install @sparticuz/chromium puppeteer-core

Problems:

Missing fonts, emoji, some rendering features
Maintenance burden when Chromium updates
Still ~100 MB with Puppeteer, leaving little room for your code
Cold starts still 5-10 seconds

3. Use Container Images

Lambda supports container images up to 10 GB.

Problems:

ECR costs and complexity
Slower cold starts (pulling larger images)
More infrastructure to manage
Still dealing with Chrome memory issues

The Simpler Solution: Don't Run Chrome in Lambda

Instead of fighting Lambda limits, use an API that runs Chrome elsewhere:

// Your Lambda function - simple HTTP call, no Chrome needed
const response = await fetch("https://api.riddledc.com/v1/run", {
  method: "POST",
  headers: {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json"
  },
  body: JSON.stringify({ url: "https://example.com" })
});

const screenshot = await response.arrayBuffer();
// Done. No Chromium. No layers. No size limits.

0 MBAdded to your deployment

~3sScreenshot returned

$0.004Per screenshot

Time Comparison

Approach	Setup Time	Cold Start	Ongoing Maintenance
Puppeteer + Layers	4-8 hours	8-15 seconds	Monthly updates
@sparticuz/chromium	2-4 hours	5-10 seconds	Version tracking
Container Image	4-6 hours	10-20 seconds	ECR + Image builds
Riddle API	5 minutes	0 (no Chrome)	None

Skip the Lambda Chrome Struggle

Get screenshots without managing Chromium infrastructure.

Get API Key View Docs