Which CI/CD systems does SBOX integrate with?

Jenkins, GitLab CI, GitHub Actions, and Azure DevOps out of the box.

Can SBOX notify Slack, Teams, or Jira?

Yes. Native integrations with Slack, Microsoft Teams, and Jira for test failure notifications, run summaries, and ticket creation.

Does SBOX require a specific LLM provider?

No. SBOX is bring-your-own-LLM. Customers connect their existing OpenAI, Anthropic, Azure OpenAI, or self-hosted model.

Where do customer LLM API keys live?

Customer-controlled keys are stored inside the customer's SBOX deployment, never on Element34 infrastructure.

Platform · Integrations & more

Integrations and more. Fits your stack. Runs in your perimeter.

Selenium Box (SBOX) plugs into the CI/CD, notification, and AI providers your team already uses. CI/CD via Jenkins, GitLab, GitHub Actions, Azure DevOps. Notifications to Slack, MS Teams, Jira. AI inference through your own LLM. Nothing leaves your perimeter.

Talk to solutions Get a quote Jump to roadmap →

CI/CD via your scheduler
Notifications without webhooks to vendor
BYO LLM, no shared inference

Section 01

CI/CD integrations

Jenkins, GitLab, GitHub Actions, Azure DevOps. Or any CI provider that calls a URL on a schedule.

Jump →

Section 02

Notifications

Slack, Microsoft Teams, Jira. Per-project webhook mapping.

Jump →

Section 03

AI capabilities

Five AI-native modules. BYO LLM. Private inference. No public egress.

Jump →

Section 04

On the roadmap

API testing, visual testing, test case management, accessibility testing.

Jump →

CI/CD Integrations

CI/CD that calls SBOX. Your pipeline, your scheduler.

SBOX exposes a schedule API your CI provider calls on the cadence you set. Jenkins, GitLab CI, GitHub Actions, Azure DevOps, or any CI provider that can call a URL on a schedule. No vendor plugin to install. No marketplace lock-in.

Jenkins

Your Jenkins job calls the SBOX schedule API at the cadence you set. Tests run on the private grid. Results land in customer storage.

GitLab CI

Your .gitlab-ci.yml stage calls the SBOX schedule API. Runners stay on your runners, test execution stays on SBOX.

GitHub Actions

A workflow step calls the SBOX schedule API. The action runner stays in your account. Test execution stays on SBOX.

Azure DevOps

A pipeline task calls the SBOX schedule API. Repos, builds, and approvals stay in Azure DevOps. Test execution stays on SBOX.

How it works

1

Create a schedule in SBOX

Get a schedule ID for your suite, project, or release branch.

2

Configure your CI provider

Make an API call to SBOX at the desired time. Any CI runner can do it.

3

Run tests at your concurrency

Hub queues, Executors run, results land in customer storage.

Any CI provider

If your CI provider can call a URL on a schedule, SBOX runs with it.

CircleCI, TeamCity, Bamboo, Bitbucket Pipelines, custom internal CI, all supported through the same schedule API.

CircleCI TeamCity Bamboo Bitbucket Pipelines Internal CI

Notifications

Notifications where your team already looks. Slack, MS Teams, Jira.

SBOX delivers test results, build summaries, and daily readiness into the channels your team already watches. Per-project webhook mapping means each project can route to its own Slack channel, MS Teams channel, or Jira project.

Slack

Build analytics alerts and daily build summaries land in the Slack channel each project owns.

Microsoft Teams

Build analytics alerts and daily build summaries deliver to the Teams channel that owns the work.

Jira

Failures and build summaries route to the Jira project that owns the workflow. Per-project mapping configurable.

Build analytics alert

Triggered per build. Each alert deep-links to the sessions page so the team can jump into the run that failed and triage from there.

Daily build summary

Delivered at the time you configure. Covers all runs from the previous 24 hours, summarized in one message.

AI capabilities

AI in every layer. Never your data leaving.

Element34 ships five AI capabilities into SBOX. Each one runs inside your tenant. Each one calls your model, not a vendor's. Each one writes to your audit trail.

Your tenant boundary

AI authoring

Studio

Prompt in plain English. Ship Selenium.

Explore →

Self-healing locators

Auto Heal

Resilient locators that survive UI churn.

Explore →

Root cause AI

Automated RCA

Cluster failures, summarize root cause, propose fixes.

Explore →

Quality intelligence

Pulse Report

AI-driven site health with prioritized opportunities.

Explore →

Customer-owned inference

BYO LLM

OpenAI, Anthropic, Azure OpenAI, or self-hosted.

Bring-your-own LLM Private inference No public egress Customer-controlled keys

On the roadmap

On the roadmap. What is coming next to SBOX.

Four capabilities in active development. Talk to your account team for delivery timelines.

✦Coming

API Testing

AI-generated API docs and AI-native API tests, scheduled alongside browser runs.

✦Coming

Visual Testing

AI-native visual diffing that ignores intentional change and surfaces real regressions.

✦Coming

Test Case Management

Native TCM linking manual cases to automated runs and tracking release readiness.

✦Coming

Accessibility Testing

WCAG-aligned a11y checks woven into SBOX runs, with AI-prioritized remediation suggestions.

FAQ

Integrations and AI, answered straight.

Ten questions buyers actually ask before signing. Answered the way we would in the room.

Which CI/CD providers does SBOX integrate with?

Jenkins, GitLab CI, GitHub Actions, and Azure DevOps are first-class. Any CI provider that can call a URL on a schedule (CircleCI, TeamCity, Bamboo, Bitbucket Pipelines, custom internal CI) works through the same schedule API.

Is there a Jenkins or GitHub marketplace plugin?

No marketplace plugin. The integration is a schedule API call. That deliberately avoids vendor plugin lock-in, and it keeps your CI provider upgrades from breaking SBOX runs.

How does our CI call SBOX?

You create a schedule in SBOX, get a schedule ID, and your CI runner calls the SBOX schedule API at the cadence you set. Tests execute on the SBOX private grid. Results land in your customer storage.

What notification destinations are supported?

Slack, Microsoft Teams, and Jira are supported today. PagerDuty and Opsgenie are on the roadmap. Two notification types: per-build analytics alert and daily build summary.

Can each project route notifications to its own Slack channel?

Yes. Each project can map to its own Slack channel, Microsoft Teams channel, and Jira project. Per-project webhook mapping is configured in the SBOX project settings.

Which LLM providers can we bring?

Enterprise OpenAI, Anthropic, Azure OpenAI, and AWS Bedrock are supported as customer-supplied keys. You bring your keys, SBOX uses your provider, your billing, your audit logs. No SBOX-hosted shared LLM by default.

Are prompts visible to us?

Yes. Prompts SBOX sends to your LLM provider are visible in your provider's audit log. SBOX makes the prompt content available to you on request inside the platform for review.

What is on the roadmap?

API testing, visual testing, test case management, and accessibility testing are on the roadmap. Direct Studio commit to GitHub or GitLab branches is on the roadmap. PagerDuty and Opsgenie alerting are on the roadmap.

When will roadmap items ship?

We do not commit to dates in marketing material. In conversation with a solutions engineer we tell you what is in dev, what is in pilot, what is in private preview, and what is committed for your contract window.

Do integrations and AI work the same way across Private Cloud, VPC, and Managed Private Cloud?

Yes. CI/CD scheduling, notification routing, and BYO LLM all work identically across the three deployment modes. What changes between modes is where the compute lives, not how integrations behave.

Talk to a solutions engineer about your stack.

We will scope CI/CD wiring, notification routing, and BYO LLM provisioning during the deployment conversation. Not a marketing page exercise.

Book a demo Get a quote

Compare deployments →