PROVIDER - CDN native image optimization

Fastly Image Optimizer

Fastly Image Optimizer builds image transforms into the CDN: AVIF, WebP and JPEG XL negotiation, URL driven sizing and cropping, smart focal cropping, all composable with VCL edge logic. A good fit for teams already on Fastly CDN, while the CDN binding and configuration depth are the gates.

Media Optimization United States 2026.11.20
4pts

Detailed Notes

Pros 5 items

Runs as a native Fastly CDN capability, so transforms happen at the edge without 1 extra image proxy
Negotiates 3 or more modern formats including AVIF, WebP and JPEG XL per client
URL parameters cover size, crop, quality and format, turning 1 source image into many variants
Smart cropping with focal points keeps subjects intact across 2 or more aspect ratios
Transforms share the CDN configuration, so VCL and edge logic can specialise 1 class of rule

Cons 4 items

Tightly bound to the Fastly CDN, which raises onboarding cost for teams on other networks
The parameter set couples with CDN configuration and needs edge configuration experience
Billing follows CDN traffic and requests, so budgets must be judged inside that model
Documentation is engineering oriented, which slows down non technical teams

Service Overview

Fastly Image Optimizer is Fastly's image optimization product from the United States, part of the CDN native branch of image and media processing. Rather than an image proxy on its own domain, it is a transform layer at the CDN edge that shares cache and VCL configuration.

The shape buys one less hop: a request lands at a Fastly edge node, gets transformed and cached, and never bounces to a separate processing service and back. The cost is a hard CDN dependency and configuration that assumes edge engineering skill.

Key Strengths

  • Inside the edge: transforms execute on CDN nodes instead of an extra origin hop
  • Format negotiation: AVIF, WebP and JPEG XL chosen per client
  • Parameterised transforms: size, quality, crop and format expressed in the URL
  • Smart cropping: focal and subject aware crops for multiple aspect ratios
  • VCL composability: edge logic can conditionally transform per business rule

Product Ecosystem

Platform capabilities

Edge image transforms, format negotiation, smart cropping, cache and purge, custom VCL logic, plus real-time logging and traffic analytics.

Integrations

Images are requested through a Fastly service domain and compose with existing CDN configuration, origins and edge compute, which suits in-house front ends and media layers.

Typical scenarios

Commerce and media sites already on Fastly, image pipelines that need edge rules, and product imagery adapted to several aspect ratios.

Limitations

  • CDN binding: hard to adopt without Fastly in place
  • Engineering depth: parameters and VCL need edge experience
  • Billing model: charges follow CDN traffic and requests
  • Documentation: engineering oriented and slow for non technical roles

Use Cases

  • Teams already on Fastly CDN (★★★★☆): capability comes with the platform
  • Edge logic customisation (★★★★☆): VCL expresses complex rules
  • Non Fastly users (★★☆☆☆): weak reason to switch
  • Small site teams (★★★☆☆): threshold exceeds the need

Pricing

Item Billing model Notes
Image optimization Billed with CDN usage Tied to traffic and requests
CDN traffic By region and volume Judge against the whole CDN budget
Enterprise Contract based Includes support and SLA

FAQ

Do I need a separate image service?

No, the capability runs on Fastly edge nodes.

Which formats are supported?

Modern formats such as AVIF, WebP and JPEG XL with per-client fallback.

Is smart cropping available?

Yes, focal and subject aware cropping supports several output ratios.

How does it compare with a self hosted service?

It removes a service and an origin hop, at the price of configuration depth.

Code Examples

<!-- Request a size and format through query parameters; the edge does the work -->
<img src="https://images.example.com/p/hero.jpg?width=1200&format=auto" alt="hero">
# Confirm the content type and cache status returned by the edge
curl -sI "https://images.example.com/p/hero.jpg?width=1200&format=auto" | grep -Ei 'content-type|cache-status'

Community & Resources

Visit Fastly Image Optimizer Website → https://www.fastly.com/products/image-optimization
Visit Website ↗