TestSprite Inc., a platform that offers end-to-end software testing using fully autonomous artificial intelligence, today announced early access to its platform today.

The company’s platform provides an AI-powered quality assurance solution that fully automates the process of testing back-end and front-end systems with minimal input from the user by using the context of the services, documents and custom instructions.

Traditional testing can be tedious and slow, requiring the generation of numerous test cases for both the user interface on the front-end and numerous test cases for code on the back-end. It is often manual and difficult to scale. The addition of AI co-pilots has eased this work considerably but still requires a noticeable degree of human assistance along the way.

TestSprite looks to take the next step and have AI learn everything it possibly can from all the context that is available about software up front and do most of the work. It can learn and understand test objects, inspect software and interpret documentation. From there, the AI automatically drafts test plans in natural language and provides them for review prior to test code generation. That means humans remain in the loop if software engineers want to customize any part of test generation.

Finally, complete testing scripts are generated based on the approved plans, scheduled and executed. Should any tests fail, the AI interprets the failures and assists with debugging the root causes using learned knowledge from software code and documentation.

“We are the first to introduce a fully automated process powered by an AI testing agent,” co-founder and Chief Executive Yunhao Jiao told SiliconANGLE. “Unlike other tools, TestSprite’s AI can independently generate testing plans, supporting over 10 types of test categories, and implement code for multiple test cases simultaneously — no manual input needed.”

TestSprite covers all the bases from user interface elements to application programming interface calls, including authorization, functionality, performance and more. It will even attempt to uncover potential edge cases based on the software code itself and the documentation. It will also auto-patch and self-fix issues in test code, while offering human review of any changes.

“One major challenge is that as more developers rely on AI co-pilots to help them write code, they become less familiar with what’s actually in their codebase,” Yunhou said. “This raises a critical question: Can they truly be confident in code that’s primarily generated by AI? Manually testing code that wasn’t entirely created by them can be daunting. That’s why using AI to validate AI-generated code becomes essential.”

The company said the software integrates with tools such as GitHub and works with common development processes such as continuous development and integration workflows. It has a wide range of support for development environments using languages such as Python and JavaScript with a focus on major cloud platforms.

TestSprite also plans to roll out test case generation for large language models and AI agents in the future. This would allow customers to use the platform to automatically generate test and run test cases for their AI-enabled applications, models and autonomous agents across broad ranges of issues. Looking ahead, Yunhao said the company also plans enhance the AI capabilities to support more programming languages and testing scenarios, including mobile apps.

The waitlist for the platform’s early access program is available for coders, developers and small businesses interested in AI-driven testing and AQ is open now at the company’s website on a first-come, first-served basis. TestSprite said the platform is targeted to launch into general availability in late October.

Image: Pixabay, TestSprite