Rust language bindings for Microsoft Playwright — the industry standard for cross-browser end-to-end testing.
Status: Pre-1.0, API stabilizing. See coverage for the path to v1.0.
This README describes the latest published release on crates.io. For changes on
mainthat haven't been released yet (new features, breaking changes, bug fixes), seecrates/playwright/CHANGELOG.mdunder[Unreleased].
Read our WHY.md to understand the vision, timing, and philosophy behind this project.
TL;DR: Rust is emerging as a serious web development language, with frameworks like Axum and Actix gaining traction. AI coding assistants are making Rust accessible to more developers. Test-Driven Development is experiencing a renaissance as the optimal way to work with AI agents. These trends are converging now, and they need production-quality E2E testing. playwright-rust fills that gap by bringing Playwright's industry-leading browser automation to the Rust ecosystem.
See Development Roadmap for plans and status of the development approach for playwright-rust.
Goal: Build this library to a production-quality state for broad adoption as @playwright/rust or playwright-rs. Provide official-quality Rust bindings for Microsoft Playwright, following the same architecture as playwright-python, playwright-java, and playwright-dotnet.
The API matches Playwright's cross-language conventions — if you know playwright-python, you know playwright-rust:
| Python | Rust |
|---|---|
from playwright.sync_api import sync_playwright
with sync_playwright() as p:
browser = p.chromium.launch()
page = browser.new_page()
page.goto("https://example.com")
# Locator with auto-waiting
heading = page.locator("h1")
assert heading.text_content() == "Example Domain"
# Response body access
resp = page.goto("https://api.example.com/data")
data = resp.json()
browser.close() |
use playwright_rs::Playwright;
let pw = Playwright::launch().await?;
let browser = pw.chromium().launch().await?;
let page = browser.new_page().await?;
page.goto("https://example.com", None).await?;
// Locator with auto-waiting
let heading = page.locator("h1").await;
assert_eq!(heading.text_content().await?, Some("Example Domain".into()));
// Response body access
let resp = page.goto("https://api.example.com/data", None).await?.unwrap();
let data: serde_json::Value = resp.json().await?;
browser.close().await?; |
Full Python API parity + agent integration. All Playwright Python
classes and methods are implemented, plus Browser::bind() /
Browser::unbind() (Playwright 1.59) for exposing a Rust-launched browser
to external clients like @playwright/mcp, the Playwright CLI, or
third-party agent tooling.
The remaining path to v1.0 is multi-month dogfooding, API polish, and performance tuning rather than new surface area. See the v1.0 gap analysis for the detailed state of each class.
playwright-rust follows Microsoft's proven architecture for language bindings:
┌──────────────────────────────────────────────┐
│ playwright-rs (Rust API) │
│ - High-level, idiomatic Rust API │
│ - Async/await with tokio │
│ - Type-safe bindings │
└─────────────────────┬────────────────────────┘
│ JSON-RPC over stdio
┌─────────────────────▼────────────────────────┐
│ Playwright Server (Node.js/TypeScript) │
│ - Browser automation logic │
│ - Cross-browser protocol abstraction │
│ - Maintained by Microsoft Playwright team │
└─────────────────────┬────────────────────────┘
│ Native protocols
┌─────────────┼─────────────┐
▼ ▼ ▼
Chromium Firefox WebKit
This means:
- ✅ Full feature parity with Playwright (JS/Python/Java/.NET)
- ✅ Cross-browser support (Chromium, Firefox, WebKit)
- ✅ Automatic updates when Playwright server updates
- ✅ Minimal maintenance - protocols handled by Microsoft's server
- ✅ Production-tested architecture used by millions
Following Playwright's cross-language consistency:
- Match Playwright API exactly - Same method names, same semantics
- Idiomatic Rust - Use Result, async/await, builder patterns where appropriate
- Type safety - Leverage Rust's type system for compile-time safety
- Auto-waiting - Built-in smart waits like other Playwright implementations
- Testing-first - Designed for reliable end-to-end testing
Add to your Cargo.toml:
[dependencies]
playwright-rs = "0.16" # Auto-updates to latest 0.16.x
tokio = { version = "1", features = ["full"] }The default-on macros feature re-exports the
locator!() compile-time
selector macro. The default-on ring feature selects the crypto backend for
driver downloads (and for rustls WebSocket connections); use aws-lc instead
when AWS-LC is required:
playwright-rs = { version = "0.16", default-features = false, features = [
"aws-lc",
"native-tls",
"macros",
] }Disabling default features requires selecting either ring or aws-lc.
If both are enabled through feature unification, the downloader uses AWS-LC;
disable default features as above to remove ring from the dependency graph.
Other opt-in features are cli (installer binary, see below) and
screenshot-diff (pixel-diff assertions). For programmatic trace-zip
inspection (CI bots, agent feedback loops), add
playwright-rs-trace as a
[dev-dependencies] entry.
See the CHANGELOG for version history and features.
Browsers must be installed before use. Install once, then run tests as many times as needed.
The library bundles a specific Playwright driver, and each driver release
expects matching browser builds. Install browsers through the crate itself
so the match is guaranteed and the browser version rides Cargo.lock:
copy
examples/install-browsers.rs
into your project's examples/ directory. (It needs tokio declared with
the macros and rt-multi-thread features.) Then:
# Install all browsers
cargo run --example install-browsers
# Or install specific browsers (in a workspace, add -p <your-package>)
cargo run --example install-browsers -- chromium firefox webkitOn Linux, required system libraries install automatically alongside the
browsers (the driver may invoke sudo for them). Pass --with-deps to
force that on other platforms; it is a no-op on macOS and needs elevation
on Windows.
In CI/CD: Add this to your GitHub Actions workflow:
- name: Install Playwright browsers
run: cargo run --example install-browsers -- chromium firefox webkitWhen dependabot bumps playwright-rs, the crate, driver, and browsers now
move together with no workflow edit. Never hardcode a Playwright version
in a workflow run: step or a package.json: dependabot can't see the
former, and it bumps the latter on npm's cadence, not the crate's. The
driver ships its own Node.js runtime, so the workflow needs no setup-node
step. For setup scripts and Docker builds, call
install_browsers
directly.
Outside a Cargo project: the cli feature provides an installer binary
(cargo install playwright-rs --features cli, then playwright-rs install). Note that cargo install builds its own copy of the crate,
which is a second version to keep in sync with your project's lockfile.
npx playwright@<version> install also works when <version> equals the
crate's bundled driver (playwright_rs::PLAYWRIGHT_VERSION, shown in the
badge above), but it goes stale silently when the crate bumps; in a Cargo
project, prefer the example approach.
What happens if I don't install browsers? You'll get an error message with install commands when trying to launch a browser.
- Rust 1.88+
- tokio async runtime
No system Node.js is needed for the normal build: the build script
downloads the pinned Playwright driver together with its own Node runtime.
(With PLAYWRIGHT_SKIP_DRIVER_DOWNLOAD set, driver resolution falls back
to PLAYWRIGHT_DRIVER_PATH or an npm-installed playwright, which do need
one.)
# Clone repository
git clone https://github.com/YOUR_USERNAME/playwright-rust.git
cd playwright-rust
# Install pre-commit hooks
pip install pre-commit
pre-commit install
# Build
cargo buildAfter building, install browsers with the in-repo example:
cargo run --package playwright-rs --example install-browsers -- chromium firefox webkitThe build script automatically downloads the Playwright driver into the
build directory ($OUT_DIR/playwright-driver); set
PLAYWRIGHT_DRIVER_CACHE_DIR to relocate it to a stable path that
survives cargo clean, which is what CI does. CI handles browser
installation automatically - see .github/workflows/test.yml.
Platform Support: ✅ Windows, macOS, Linux
Known limitation: WebKit launch_persistent_context() fails on native
Windows with "Initial load failed" — this is an upstream Playwright issue
(microsoft/playwright#36936,
also tracked as playwright-rust #39).
Microsoft is building a channel: "webkit-wsl" replacement
(microsoft/playwright#37036).
Chromium and Firefox persistent contexts work on all platforms. Non-persistent
WebKit (browser.new_context()) works on Windows. Use WSL or macOS/Linux for
WebKit persistent contexts.
This project uses cargo-nextest. Install once: cargo install cargo-nextest
cargo nextest run # All tests
cargo nextest run -p playwright-rs --lib # Unit tests only (~2s, no browsers)
cargo nextest run -p playwright-rs -E 'test(locator)' # Pattern match
cargo test --doc --workspace # Doc-tests (compile-checked)See examples/ for usage examples.
cargo run --package playwright-rs --example basicSee the source line that failed. When ? propagates an Error out of
your test, you see the message but no Rust source location. Use
anyhow for tests and run with RUST_BACKTRACE=1:
use anyhow::{Context, Result};
#[tokio::test]
async fn my_test() -> Result<()> {
// ...
let content = heading.text_content().await.context("read heading")?;
// ...
Ok(())
}Run as RUST_BACKTRACE=1 cargo nextest run (or cargo test). The backtrace
points at the failing ?, and .context("...") adds breadcrumbs to the
error chain. This matches how playwright-java/dotnet rely on the test
runner's stack trace rather than baking source locations into the library.
Save a Playwright trace when a test fails. Rust has no async Drop,
so trace cleanup is explicit. Capture the test result, then run cleanup
unconditionally and pass the trace path only on failure:
let result = run_test_body(&context).await;
let stop_opts = result.is_err().then(|| TracingStopOptions::default().path("trace.zip"));
let _ = tracing.stop(stop_opts).await;
let _ = browser.close().await;
result?;See examples/trace_on_failure.rs
for a runnable end-to-end example. Open the resulting trace.zip at
https://trace.playwright.dev.
This repo ships an Agent Skill so your agent
writes against this crate's actual API model instead of guessing from
generic Playwright knowledge. It covers adding the dependency and
installing browsers, the object model, the builder and locator!()
conventions, auto-wait semantics, and capturing a trace when something
fails.
npx skills add padamson/playwright-rustWorks with Claude Code, Codex, Cursor, and any other compatible agent.
Claude Code can also install it as a plugin, which tracks this repo:
/plugin marketplace add padamson/playwright-rust
/plugin install playwright-rs@playwright-rustThird-party marketplaces don't refresh on their own, so pick up a later version with both commands, catalog first:
/plugin marketplace update playwright-rust
/plugin update playwright-rs@playwright-rustThe skill lives at
skills/playwright-rs-usage/. It points back
at docs.rs and the
examples for the API surface itself, being a
thin "what to reach for, what to avoid" overlay rather than a duplicate of
the API reference. A build gate keeps it honest: its code compiles against
the real crate, and it cannot omit a cargo feature or browser engine the
crate exposes.
This project aims for production-quality Rust bindings matching Playwright's standards. Contributions should:
- Follow Playwright API conventions
- Include comprehensive tests
- Maintain type safety
- Document public APIs with examples
- Pass CI checks (fmt, clippy, tests)
Apache-2.0 (same as Microsoft Playwright)
- Microsoft Playwright Team - For the amazing browser automation framework
- playwright-python - API design reference