About this project

Coding agents have become the platform developers build workflows on top of. But building for agents feels too similar to building for the web in the early aughts: fragmented support, unclear standards, and lots of wondering "does this actually work on all browsers and devices?" Agents shouldn't follow the same pattern. We're better than that.

Enter: Can I Use Agents?

A tool to properly check compatibility across agents without having to manually check docs. Heavily inspired by caniuse.com for browsers, caniuseagents tracks feature support across coding agents so you can make informed decisions about which tools fit your workflow and keep up with the growing ecosystem around AI agents.

What we track

Each feature page shows a compatibility matrix with historical version data. In addition to seeing whether something is supported today, you can see when support was introduced and how it evolved over time. Features are generally organized into categories:

Which agents

We currently track Claude Code , Cline , Codex CLI , Copilot , Cursor , and Windsurf . These are the agents developers are actively building on top of — not just using for chat, but integrating into CI pipelines, editor workflows, and automated toolchains.

Why this matters

As agents become part of the modern developer experience, compatibility becomes infrastructure. Agent features like lifecycle hooks and specific protocol support directly affect how tools behave, integrate, and fail. caniuseagents exists to bring clarity to an otherwise chaotic agent landscape.

Contributing

The data behind this site is open source. If you spot something outdated, want to add a feature, or have support data for an agent, contributions are welcome on GitHub.