Cerberus Testing

The Open Source Test Automation Platform

  • 0
    • Start Now

    AI Testing Self-Hosted: Setup Guide in Under 30 Minutes

    By Antoine CraskePosted on Posted in AI Testing, TutorialsTagged , , , ,

    Setting up AI testing self-hosted infrastructure used to take days of configuration, license negotiations, and infrastructure planning. With the rise of open-source AI-powered QA platforms, you can now deploy a production-grade test automation environment on your own servers in under 30 minutes, without sending a single test artifact to a third-party SaaS.

    This step-by-step tutorial walks you through deploying Cerberus Testing, an open-source AI test automation platform, on your own infrastructure using Docker. By the end, you will have a running instance, your first AI-generated test case, and a clear path to integrate continuous testing in your CI/CD pipeline.

    Why Choose AI Testing Self-Hosted Over SaaS?

    Before diving into the setup, it is worth clarifying why a self-hosted AI testing platform is becoming the preferred choice for QA teams in 2025:

    • Data sovereignty: Test data, screenshots, and application traffic stay inside your network.
    • No vendor lock-in: You own your test repository, execution logs, and AI models.
    • Predictable costs: No per-user or per-execution fees that explode with scale.
    • Compliance-ready: Easier to meet GDPR, HIPAA, or ISO 27001 requirements.
    • Customization: Plug in your own AI models, internal tools, and authentication providers.

    For regulated industries such as banking, healthcare, and public services, AI testing without SaaS is often the only viable path. Open-source self-hosted platforms close that gap.

    System Requirements for Self-Hosted AI Test Automation

    Cerberus Testing is designed to run on commodity hardware. For a development or proof-of-concept environment, the following minimal configuration is enough:

    • OS: Linux (Ubuntu 22.04+, Debian 12, RHEL 9) or macOS for local testing
    • CPU: 4 vCPUs (8 recommended for parallel execution)
    • RAM: 8 GB minimum, 16 GB recommended when running AI features
    • Disk: 20 GB free SSD storage
    • Docker: Docker Engine 24+ and Docker Compose v2
    • Network: Outbound access to pull container images (or your private registry)

    For production deployments running large regression test automation suites, scale horizontally using Kubernetes or a managed orchestration tool. The architecture supports stateless executors and a central MariaDB-backed core.

    Step 1: Install Docker and Docker Compose

    On Ubuntu, install the Docker Engine with the official script:

    • Run curl -fsSL https://get.docker.com | sh
    • Add your user to the docker group: sudo usermod -aG docker $USER
    • Log out and back in, then verify with docker --version and docker compose version

    For other platforms, follow the official Docker installation guide. Confirm Docker is running with docker run hello-world before proceeding.

    Step 2: Pull the Cerberus Testing Docker Stack

    Cerberus provides a ready-to-use Docker Compose configuration that bundles the core application, database, and execution engine. Create a working directory and download the stack:

    • mkdir cerberus-ai && cd cerberus-ai
    • curl -O https://raw.githubusercontent.com/cerberustesting/cerberus-source/main/docker-compose.yml
    • Review the file and adjust environment variables (database password, exposed ports, AI provider keys if any)

    The stack includes the Cerberus core, a MariaDB instance, a Selenium Grid for browser-based testing, and the AI engine responsible for smart selectors and self-healing tests.

    Step 3: Launch Your Self-Hosted AI Testing Platform

    Start the entire environment with a single command:

    • docker compose up -d
    • Wait approximately 2 to 3 minutes for the database migration and initial bootstrap
    • Tail the logs to confirm a clean startup: docker compose logs -f cerberus

    Once the logs show Server startup completed, open your browser at http://localhost:8080/Cerberus. Log in with the default administrator credentials, then immediately change the password from the user management screen.

    Verifying the AI Engine

    Navigate to the Engine Settings page and confirm that the AI service is reachable. A green status indicator means smart selector generation, test case suggestions, and self-healing are active. If you plan to connect an external LLM provider for richer AI capabilities, configure the endpoint and API key here.

    Step 4: Create Your First AI-Generated Test Case

    This is where the value of an AI-powered QA automation platform becomes tangible. Instead of scripting selectors manually, let the AI do the heavy lifting:

    • Click Test Case > New Test Case in the main menu
    • Define the application URL (for example, your staging environment)
    • Describe the scenario in plain English: Open the login page, enter valid credentials, submit, and verify the dashboard loads
    • Click Generate with AI

    Cerberus parses your description, analyzes the target page, and produces a codeless test case with resilient selectors. You can then refine the steps using the visual editor, no programming required. This codeless test automation flow is ideal for agile teams where developers, testers, and product owners collaborate on the same artifact.

    Run and Inspect the Execution

    Trigger the execution from the test case view. Cerberus dispatches the test to the local Selenium Grid, records each step with screenshots, and the AI engine validates that selectors remained stable. If a selector drifts, the self-healing module proposes an updated locator automatically.

    Step 5: Connect to Your CI/CD Pipeline

    A self-hosted AI testing platform delivers maximum value when integrated into continuous testing CI/CD workflows. Cerberus exposes a REST API to trigger campaigns from any pipeline:

    • GitHub Actions: Add a step calling POST /api/RunTestCaseV001 with your campaign ID
    • GitLab CI: Use a curl command in a dedicated regression stage
    • Jenkins: Use the HTTP Request plugin to launch test campaigns on every release candidate

    Reports are available as JSON, JUnit XML, or directly in the Cerberus dashboard, ready to feed your release gates.

    Next Steps to Scale Your AI Test Automation

    You now have a fully functional, open-source, self-hosted AI testing platform running in your environment. To go further:

    • Add browser nodes to the Selenium Grid for parallel regression testing automation
    • Connect SSO via LDAP or SAML for enterprise authentication
    • Define test data services to feed dynamic scenarios
    • Explore the API test automation modules for backend coverage

    Self-hosting an AI-powered test automation platform is no longer reserved for large engineering teams. With Docker, sensible defaults, and an open-source license, any QA team can own its testing stack end to end.

    Ready to dive deeper? Explore our documentation, star the project on GitHub, and join the community Slack to share your setup experience.

    AI Testing Self-Hosted: Setup Guide in Under 30 Minutes
    Scroll to top