Insights from building the Canva Developers Platform to empower 185 million designers

How do you build a public API for 185 million users without slowing internal innovation? Canva's answer is a crucial 'anti-corruption layer'.

#1about 3 minutes

Introducing the Canva Developers Platform and SDK

Canva launched its developer platform and SDK to empower a community of developers to meet the diverse design needs of its 185 million users.

#2about 3 minutes

Understanding the architecture of the Canva Apps SDK

The SDK uses a sandboxed iframe and a message bus to communicate with an API controller, which acts as an anti-corruption layer to decouple public APIs from internal product code.

#3about 3 minutes

Using internal hackathons and feedback to build APIs

Canva uses a phased approach including internal releases and hackathons for dogfooding to gather feedback and improve APIs before a public launch.

#4about 4 minutes

Starting small with a pragmatic engineering approach

The team initially prioritized speed by using manual processes and simple infrastructure, allowing them to focus on delivering APIs before investing in automated pipelines.

#5about 3 minutes

Prioritizing API development with app archetypes

Canva defined three core app archetypes—creation, content, and AI-powered—to ruthlessly prioritize which APIs to build next and align the engineering teams.

#6about 4 minutes

Following four core principles for outstanding API design

APIs are designed as a long-term commitment guided by four principles: simplicity, safety, evolvability, and consistency to ensure a great developer experience.

#7about 3 minutes

Ensuring API consistency with a design working group

An API design working group, run sheets, and a shared engineering handbook help maintain consistency and alignment across independent teams, avoiding the pitfalls of Conway's Law.

#8about 2 minutes

Key lessons learned from building a developer platform

The journey highlights the importance of dogfooding, pragmatic excellence, avoiding poor APIs, and writing things down to gain clarity and scale effectively.