What is Silverstream?
Silverstream is a cloud infrastructure enabling enterprise to build, scale and monitor web browsing agents.
Silverstream QA testing solution, powered by Silverstream web agents, allows full automation of E2E frontend testing of web application and website.
Why Silverstream QA agentic solution?
Silverstream QA agents leaverage web-understanding capability to interact autonomously with web pages. Instead of generating static Playwright scripts,the agents interact with the browser via a live instance of Playwright.
This eliminates manual intervention, as there is no need for writing, updating, or maintaining test scripts—agents dynamically adapt and execute tests on their own.
How to test with Silverstream QA agentic solution?
Install the silverriver library through pip. The dependencies included are Playwright and Chromium.
pip install silverriver==0.1.37b9
Demonstrate the workflow you want to test on an an instrumented Chromium session that record your actions:
silverriver record URL -o filename
Close the browser to save the trace and upload the file at dashboard.silverstream.ai.
The dashboard generates a user story from your recording, including context, steps, and acceptance criteria.
When you run the test, the agent executes it in headless Chromium and provides comprehensive logs viewable at trace.playwright.dev
* Success: acceptance criteria are met
* Failure: task not completed, there is a bug.
* Skipped: the test is not feasible
* Error: for system-level issues (rare)
Capabilities and Limitations
Good Fit (Capabilities) | Bad Fit (Limitation) |
Parallel workflow tests | Multi browser tests |
Dynamic content (functional test) | Unit and performance test |
Web application tests in Chromium | Mobile Application Tests |
How does the agent “understand” the browser? Does it rely on selectors?
The agent rely on Silverstream multi-modal web page understanding (DOM parsing, viewport screenshots, AXTree). This enables reliable interaction independent of traditional selectors, while handling dynamic content (like popups and iframes) and complex navigation patterns.
How does the agent identify when a workflow is complete?
The workflow completion heavily relies on the acceptance criteria. Make sure to check them after the generation of the user story. (e.g., reaching a thank you page after checkout). You can always check the session logs for verification.
Can I still define specific test cases?
The core test path follows your user story and its acceptance criteria. Soon, a new feature will enable automatic test generation for open-ended exploration.
How do I integrate Silverstream into my existing CI/CD pipeline?
Silverstream has a fully managed testing approach:
- No test writing or mantaining
- No infrastructure setup required
- On-demand test execution
How can I debug with Silverstream?
At the end of the agent run, you can download the Playwright trace of the session for debugging. More info here.
Still Have Questions?
If you have additional questions about how Silverstream works or how it fits into your current testing strategy, feel free to reach out at hello@silverstream.ai