#selenium #webdriver #chromedriver #geckodriver #automation


Thirtyfour is a Selenium / WebDriver library for Rust, for automated website UI testing. It supports the full W3C WebDriver spec

65 releases (26 breaking)

0.27.2 Oct 11, 2021
0.27.0 Aug 6, 2021
0.26.0 Jul 7, 2021
0.23.0 Mar 27, 2021
0.1.0 Nov 30, 2019

#21 in Testing

Download history 440/week @ 2021-07-02 306/week @ 2021-07-09 293/week @ 2021-07-16 147/week @ 2021-07-23 276/week @ 2021-07-30 407/week @ 2021-08-06 289/week @ 2021-08-13 155/week @ 2021-08-20 77/week @ 2021-08-27 153/week @ 2021-09-03 196/week @ 2021-09-10 270/week @ 2021-09-17 139/week @ 2021-09-24 178/week @ 2021-10-01 220/week @ 2021-10-08 151/week @ 2021-10-15

899 downloads per month
Used in 5 crates (3 directly)



Crates.io docs.rs Build Status

Thirtyfour is a Selenium / WebDriver library for Rust, for automated website UI testing.

It supports the full W3C WebDriver spec. Tested with Chrome and Firefox although any W3C-compatible WebDriver should work.

Async only (tokio and async-std runtimes supported via feature flags). For synchronous support, use the thirtyfour_sync crate instead.


  • All W3C WebDriver and WebElement methods supported
  • Async / await support (both tokio and async-std runtimes supported via feature flags)
  • Create new browser session directly via WebDriver (e.g. chromedriver)
  • Create new browser session via Selenium Standalone or Grid
  • Find elements (via all common selectors e.g. Id, Class, CSS, Tag, XPath)
  • Send keys to elements, including key-combinations
  • Execute Javascript
  • Action Chains
  • Get and set cookies
  • Switch to frame/window/element/alert
  • Shadow DOM support
  • Alert support
  • Capture / Save screenshot of browser or individual element as PNG
  • Chrome DevTools Protocol (CDP) support
  • Advanced query interface including explicit waits and various predicates

Why 'thirtyfour' ?

It is named after the atomic number for the Selenium chemical element (Se).

Feature Flags

  • tokio-runtime: (Default) Use the tokio async runtime with the reqwest http client.

  • async-std-runtime: Use the async-std runtime with the surf http client.

    NOTE: There are also additional features for enabling TLS. See the API documentation for details.


The examples assume you have a WebDriver running at localhost:4444.

You can use Selenium (see instructions below) or you can use chromedriver directly by downloading the chromedriver that matches your Chrome version, from here: https://chromedriver.chromium.org/downloads

Then run it like this:

chromedriver --port=4444

Example (async):

To run this example:

cargo run --example tokio_async
use thirtyfour::prelude::*;
use tokio;

async fn main() -> WebDriverResult<()> {
     let caps = DesiredCapabilities::chrome();
     let driver = WebDriver::new("http://localhost:4444", &caps).await?;

     // Navigate to https://wikipedia.org.
     let elem_form = driver.find_element(By::Id("search-form")).await?;

     // Find element from element.
     let elem_text = elem_form.find_element(By::Id("searchInput")).await?;

     // Type in the search terms.

     // Click the search button.
     let elem_button = elem_form.find_element(By::Css("button[type='submit']")).await?;

     // Look for header to implicitly wait for the page to load.
     assert_eq!(driver.title().await?, "Selenium - Wikipedia");
     // Always explicitly close the browser. There are no async destructors.


The browser will not close automatically

Rust does not have async destructors, which means there is no reliable way to execute an async HTTP request on Drop and wait for it to complete. This means you are in charge of closing the browser at the end of your code, via a call to WebDriver::quit() as in the above example.

If you do not call WebDriver::quit() then the browser will stay open until it is either explicitly closed later outside your code, or the session times out.

Advanced element queries


The WebDriver::query() and WebElement::query() methods return an ElementQuery struct.

Using ElementQuery, you can do things like:

let elem_text =

This will execute both queries once per poll iteration and return the first one that matches. You can also filter on one or both query branches like this:


The all() method will return an empty Vec if no elements were found. In order to return an error in this scenario, use the all_required() method instead.

ElementQuery also allows the use of custom predicates that take a &WebElement argument and return a WebDriverResult<bool>.

As noted above, the query() method is also available on WebElement structs as well for querying elements in relation to a particular element in the DOM.


The WebElement::wait_until() method returns an ElementWaiter struct.

Using ElementWaiter you can do things like this:

// You can optionally provide a nicer error message like this.
elem.wait_until().error("Timed out waiting for element to disappear").not_displayed().await?;


And so on. See the ElementWaiter docs for the full list of predicates available.

ElementWaiter also allows the use of custom predicates that take a &WebElement argument and return a WebDriverResult<bool>.

A range of pre-defined predicates are also supplied for convenience in the thirtyfour::query::conditions module.

use thirtyfour::query::conditions;


These predicates (or your own) can also be supplied as filters to ElementQuery.

Running against selenium

NOTE: To run the selenium example, start selenium (instructions below) then run:

cargo run --example selenium_example

Below you can find my recommended development environment for running selenium tests.

Essentially you need 3 main components as a minimum:

  1. Selenium standalone running on some server, usually localhost at port 4444.

    When using selenium, remember to include /wd/hub after the server address. For example, when running against localhost:4444, your selenium url will be http://localhost:4444/wd/hub

  2. The webdriver for your browser somewhere in your PATH, e.g. chromedriver (Chrome) or geckodriver (Firefox)

  3. Your code, that imports this library

If you want you can download selenium and the webdriver manually, copy the webdriver to somewhere in your path, then run selenium manually using java -jar selenium.jar.

However, this is a lot of messing around and you'll need to do it all again any time either selenium or the webdriver gets updated. A better solution is to run both selenium and webdriver in a docker container, following the instructions below.

Setting up Docker and Selenium

To install docker, see https://docs.docker.com/install/ (follow the SERVER section if you're on Linux, then look for the Community Edition)

Once you have docker installed, you can start the selenium server, as follows:

docker run --rm -d -p 4444:4444 -p 5900:5900 --name selenium-server -v /dev/shm:/dev/shm selenium/standalone-chrome-debug:3.141.59-zinc

We will use the -debug container because it starts a VNC server, allowing us to view the browser session in real time. To connect to the VNC server you will need to use a VNC viewer application. Point it at localhost:5900 (the default password is secret). If you don't want to use a VNC viewer you don't have to, but then you won't see what is happening when you run the examples.

If you want to run on Firefox instead, or customize the browser setup, see docker-selenium on GitHub for more options

Choosing between Sync and Async

The thirtyfour crate provides an async API whereas thirtyfour_sync provides sync. Which one you should use really depends on your personal preference and the nature of your application.

For a more in-depth introduction to async programming in rust, see The Rust Async Book.

All interactions with selenium involve sending requests to the selenium server, and then waiting for a response. The difference between sync and async for thirtyfour essentially comes down to what actually happens (or doesn't happen) while waiting for that response.

With the synchronous version, requests to selenium will block the calling thread until they return. While this is obviously inefficient, the code is often a little bit simpler and easier to follow. You can alleviate some of the inefficiency by using multiple threads if/when concurrency is required, but note that threads have some overhead as well. If you're thinking of spawning hundreds (or even tens) of threads, you probably want to use async instead.

In contrast, async method calls will immediately return a "future", which is a bit like a task that can be executed later (typically via .await), and will eventually complete and return a value. While it is .awaiting a response from selenium, the "future" will yield control of the current thread, allowing other "futures" to run. This allows multiple operations to be performed concurrently on the same thread, rather than just sitting and blocking the thread completely for each call to selenium. For more information on running multiple tasks concurrently, see chapter 6 of The Rust Async Book.

In short, if you need to do a lot of I/O (network requests, reading/writing files, handling network connections), and especially if you want to do I/O operations concurrently, choose async. If not, it's really up to your personal preference.

WebDriverSession Lifetime

All WebElement structs share a reference to the WebDriverSession which provdes a compile-time guarantee that no element or alert struct (for example) will outlast the browser session. This should prevent issues where something attempts to send a command to a browser session that has already been closed.

By embedding a reference to the actual WebDriverSession inside each struct such as WebElement, this also enables things such as easily adding WebElement methods that run JavaScript internally.

Running the tests for thirtyfour, including doctests

You generally only need to run the tests if you plan on contributing to the development of thirtyfour. If you just want to use the crate in your own project, you can skip this section.

Just like the examples above, the tests in this crate require a running instance of Selenium at http://localhost:4444.

If you still have the docker container running from the examples above, you will need to bring it down (otherwise the next step will complain that port 4444 is already in use):

docker stop selenium-server 

The tests also require a small web app called webappdemo that was purpose-built for testing the thirtyfour crate.

Both can be run easily using docker-compose. To install docker-compose, see https://docs.docker.com/compose/install/

Once you have docker-compose installed, you can start the required containers, as follows:

docker-compose up -d

Then, to run the tests:

cargo test -- --test-threads=1

We need to limit the tests to a single thread because the selenium server only supports 1 browser instance at a time. (You can increase this limit in the docker-compose.yml file if you want. Remember to restart the containers afterwards)

If you need to restart the docker containers:

docker-compose restart 

And finally, to remove them:

docker-compose down


This work is dual-licensed under MIT or Apache 2.0. You can choose either license if you use this work.

See the NOTICE file for more details.

SPDX-License-Identifier: MIT OR Apache-2.0


~300K SLoC