Selenium News and Updates: A Smattering of Selenium 100

At the research cutoff of October 3, 2026, Selenium’s official blog listed Selenium 4.49, released September 9, as its newest release. Other updates worth developers’ attention include a planned Java API change in 4.51, an official GitHub Container Registry mirror for Selenium Grid images, and the project’s stated direction toward WebDriver BiDi. These are separate developments: a released version, a future compatibility change, an image distribution option, and protocol work.

What’s new in Selenium?

The Selenium blog archive is the source of record for the project announcements summarized here. Its September 2026 entries cover the 4.49 release, the planned Java change for 4.51, and a documentation page about errors in code written by AI agents. The archive also lists 4.48 and 4.47 releases in August. That sequence shows recent release activity, but it does not establish a guaranteed release schedule. Read the official Selenium blog and archive. [Selenium blog]

Selenium 4.49: the latest release listed at the research cutoff

Selenium announced version 4.49 on September 9, 2026. The project’s announcement said, “Today we’re happy to announce that Selenium 4.49 has been released!” This is a release announcement, not a claim that 4.49 will remain the latest version after the research date. Check the project’s release notes and your language binding’s compatibility details before upgrading. [Selenium official blog archive]

The archive also lists 4.48 on August 27 and 4.47 on August 10, 2026. Those dates provide context for recent releases, but should not be used to predict when the next one will ship. [Selenium official blog archive]

Java users: prepare for the ExpectedCondition change in 4.51

On September 21, the Selenium project announced that, starting with Selenium 4.51, Java’s ExpectedCondition will stop implementing Guava’s Function interface. This was a planned change at the research cutoff; treat it as an upgrade compatibility item and verify the migration guidance before moving to 4.51. [Selenium official blog archive] [Selenium announcement and migration detail]

Who should review their code?

Java maintainers should inspect code that treats an ExpectedCondition as a Guava Function, passes it to APIs requiring that interface, or otherwise relies on the type relationship. Ordinary use of Selenium wait conditions may not depend on that interface, but check the actual types and compiler errors in your project rather than assuming.

Upgrade checklist

Search Java sources for ExpectedCondition and Guava Function usage, including helper methods and wrappers.
Read the Selenium migration detail linked from the announcement and identify affected call sites.
Compile and run your wait-related checks against the version you plan to adopt in a branch or pre-release validation environment.
Update adapters or method signatures that require Guava’s interface, then upgrade deliberately.

The key distinction is timing: the blog announced a change beginning with 4.51; it was not described as already present in 4.49. [Selenium official blog archive]

Selenium Grid images are mirrored to GitHub Container Registry

Selenium announced that its official Grid Docker images are mirrored to GitHub Container Registry (GHCR), describing GHCR as an official alternative to Docker Hub. Teams can consider the mirror when configuring where their Grid images are pulled from. Choose the registry that fits your organization’s image policy and deployment setup; the announcement does not establish that either registry is universally faster or cheaper. [Selenium official blog archive]

What Grid operators should check

Review image references in deployment manifests, compose files, and automation scripts.
Confirm the registry host and image tags against Selenium’s official announcement and current image documentation.
Check that your CI workers and cluster nodes can authenticate to and pull from the selected registry.
Keep the version tag explicit and follow your normal image update and rollback process.

The archive also lists a Grid 4.41.0 update and the Selenium 4.41 release in February 2026, as well as the GHCR mirror announcement in April. These are historical archive entries; use current official documentation for the image path and tag you deploy. [Selenium official blog archive]

WebDriver BiDi is the stated direction for browser protocol work

The Selenium archive says Chrome DevTools support for Firefox was removed and describes WebDriver BiDi as the way forward. It also notes that Puppeteer has moved to support WebDriver BiDi in Chromium browsers and Firefox. This indicates ecosystem direction, not complete feature parity: browser support and individual BiDi capabilities can differ by implementation. Confirm the specific browser, binding, and feature you need before changing a test architecture. [Selenium official blog archive]

How to assess a BiDi migration

List the browser-specific capabilities your tests currently use.
Check the Selenium documentation for support in your chosen browser and binding.
Run representative tests across your actual browser matrix; a protocol name alone does not guarantee identical behavior.
Keep a fallback plan for capabilities that are not yet available in a particular implementation.

Other archive updates worth knowing

AI agent generated Selenium code

A September 29, 2026 post says Selenium added a documentation page addressing recurring errors in Selenium code written by coding agents. Developers using code assistants can use that material as a review aid, while still checking generated code against the language binding’s current documentation and their application’s behavior. [Selenium official blog archive]

Chrome version 127 telemetry report

A 2025 Selenium post reported telemetry indicating over 28 million users locked to Chrome version 127.0.6533.99. This is a historical observation reported by the Selenium project about that specific browser version. It is not a current Selenium user count or an independently verified estimate. [Selenium 2025 archive]

Python 3.9 support announcement

The 2025 archive includes an announcement that Selenium’s Python WebDriver bindings were ending Python 3.9 support. Because this is an older, version-specific compatibility item, Python maintainers should consult current Selenium support documentation rather than infer present support status from that announcement alone. [Selenium 2025 archive]

Cloud browser testing and project partners

Selenium’s archive identifies BrowserStack, Sauce Labs, and TestMu AI (formerly LambdaTest) as development partners. That establishes the names in the context of Selenium project announcements; it does not establish current commercial terms, affiliate access, or the capabilities of a particular service. Teams evaluating remote browser testing should check each provider’s current documentation and terms directly. [Selenium official blog archive]

What developers should do next

For all users: read the 4.49 release information and check the current release page before selecting a version.
For Java maintainers: audit reliance on Guava Function in connection with ExpectedCondition before upgrading to 4.51.
For Grid operators: review whether the official GHCR mirror fits your pull policy and update image references only after verifying the current image path.
For browser protocol users: evaluate WebDriver BiDi support feature by feature and browser by browser.
For teams using coding agents: include Selenium’s agent-error guidance in code review, then validate the generated test against the actual application.

Or skip the browser setup

If your task is capturing a page for a report, visual reference, or downstream workflow rather than automating browser interactions, ScreenshotNeo is a website screenshot API and MCP server. It returns a PNG, JPEG, WebP, or PDF from one GET request. See the ScreenshotNeo API documentation.

curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webp

import requests

r = requests.get(
    "https://api.screenshotneo.com/v1/shot",
    params={"access_key": "YOUR_API_KEY", "url": "https://stripe.com"},
    timeout=90,
)
open("shot.webp", "wb").write(r.content)

const q = new URLSearchParams({ access_key: 'YOUR_API_KEY', url: 'https://stripe.com' });
const res = await fetch(`https://api.screenshotneo.com/v1/shot?${q}`);
if (!res.ok) throw new Error(`Screenshot request failed: ${res.status}`);
await Bun.write('shot.webp', res);

Cookie and consent banners, newsletter popups, and chat widgets are removed before the shot; each cleanup step can be turned off. Bot checks, blank pages, failed loads, timeouts, and cache hits are not billed, and response headers report the page verdict and billing status. An MCP server lets AI agents—including Claude, Cursor, and other MCP clients—take screenshots. The free plan includes 1,000 screenshots a month with no card; paid plans start at $5 for 3,000 shots. Sign up free for ScreenshotNeo.

Performance, reliability, and cost notes

For Selenium updates, avoid treating release cadence as a reliability guarantee. Pin versions in CI where reproducibility matters, review release and migration notes before upgrading, and validate against the browser versions and capabilities your tests actually use. For Grid, use the registry your infrastructure can reliably access and verify pulls in the same environment where the nodes run. These are operational recommendations; the cited announcements provide no comparative speed, uptime, or cost benchmarks.

For the screenshot API option above, ScreenshotNeo’s listed plans are Free: 1,000 shots per month; Starter: $5 for 3,000; Growth: $15 for 15,000; Pro: $39 for 60,000; Scale: $99 for 250,000; Business: $249 for 1,000,000. Yearly billing gives two months free, and every feature is available on every plan. Only clean shots are billed; the response’s X-Page-Verdict and X-Billed headers indicate outcome and billing status. [ScreenshotNeo documentation]

Troubleshooting these updates

Symptom	Likely cause	What to do
Java compile error after moving to 4.51	Code depends on `ExpectedCondition` implementing Guava `Function`.	Consult the migration detail and update affected signatures or adapters; compile and validate wait behavior.
Grid deployment cannot pull the image	Registry hostname, tag, access policy, or network configuration does not match the configured mirror.	Verify the official image reference and tag, check node connectivity and authentication, then retry the pull.
A BiDi test works in one browser but not another	Browser or implementation support differs for the particular feature.	Check support for that browser, binding, and capability; run a small representative test before broad migration.
Generated Selenium code fails intermittently or uses stale APIs	The generated code may contain recurring agent errors or assume outdated binding behavior.	Review against Selenium’s agent-related documentation and current binding docs, then reproduce with a minimal test.
Screenshot API response is not an image	The capture may have a page verdict such as a bot check, blank page, timeout, or failed load.	Inspect `X-Page-Verdict` and `X-Billed`; adjust URL or capture settings as appropriate. Failed outcomes are not billed.

FAQ

Is Selenium 4.49 still the latest version?

It was the newest release listed in the official Selenium blog at the research cutoff, October 3, 2026. Check the current archive before acting on that status.

Does the 4.51 Java change affect every ExpectedCondition use?

The announced change concerns the Guava Function interface relationship. Review code that relies on that relationship and follow Selenium’s migration detail.

Does WebDriver BiDi mean every browser has the same capabilities?

No. The archive describes BiDi as the direction forward, but support can vary by browser and feature. Verify the capability you need.

Can ScreenshotNeo replace Selenium?

It handles website screenshots and PDFs through an API or MCP server. It is not presented as a replacement for Selenium’s browser automation and interaction workflows.