Papercuts

The second major feature group is Exploratory agents, which enable autonomous discovery within your application. You simply provide the URL, and these agents roam your app like a curious user, clicking, scrolling, and uncovering hidden edge cases. Unlike deterministic agents, exploratory agents do not follow a scripted path; instead, they explore the application freely to find issues that you may not have anticipated. This is invaluable for catching regressions, broken links, misrendered components, and other subtle bugs that occur under real-world conditions. The exploratory approach simulates the behavior of a new user navigating your site without guidance, revealing weaknesses in user experience and functionality that could otherwise go unnoticed.

Additional capabilities are outlined in the Papercuts protocol, which consists of three steps: Inject, Perceive, and Report. During injection, a fleet of headless browsers is instantiated to enter the application exactly as a user would, requiring no SDKs or code changes. The perception phase involves agents that see the UI semantically, not just code, allowing them to adapt to layout shifts without breaking tests. Finally, the report phase sends an email notification with a detailed diagnostic pinpointing exactly where the error occurred. This protocol also supports authentication by allowing you to provide credentials for protected routes, ensuring that all parts of your application can be tested in production.

Overall, Papercuts operates on the philosophy that production is the only environment that matters. The workflow begins with you providing a URL and optionally any authentication credentials. The platform then deploys either deterministic or exploratory agents (depending on your plan) to interact with the application. These agent instances run in headless browsers, perceiving the UI semantically to perform actions and check for errors. When an issue is detected, the system automatically generates a comprehensive report that includes a step-by-step breakdown of the execution flow, high-resolution screenshots, network request logs, and deep agent reasoning logs. This report is immediately emailed to you, enabling quick identification and resolution of production bugs.

Concrete use cases for Papercuts include validating critical e-commerce checkout flows where a deterministic agent adds an item to the cart and completes the purchase, ensuring no transaction failures. Another scenario involves using exploratory agents to test a newly deployed feature across all pages, uncovering layout shifts or broken navigation that might affect user satisfaction. Teams can also set up recurring tests for authenticated routes, logging in with provided credentials to verify that user-specific content loads correctly. The outcome is a detailed diagnostic report that saves hours of manual debugging, allowing developers to fix issues before they impact customers. Users report catching bugs that never appeared in staging environments.

Papercuts is designed for frontend and backend developers, QA engineers, and product managers who work on web applications and need a reliable way to catch production-only bugs. It supports any web application without requiring specific frameworks or SDKs. Pricing starts with a Free tier that includes 20 actions, 3 test flows, and deterministic agents only, ideal for small teams getting started. The Pro plan at $5 per month offers 500 actions, unlimited test flows, and both deterministic and exploratory agents, with additional actions at $0.02 each. This makes Papercuts accessible for teams of all sizes. In summary, Papercuts delivers a powerful, no-integration solution for AI production testing, ensuring that your application behaves correctly in the hands of real users.