Cursor AI Agents Now Demo with Video
The AI-native code editor Cursor has launched a new feature where its AI agents demonstrate software changes with video recordings instead of text-based diffs. The update includes the ability to run and record end-to-end backend tests, such as redacting user data. In a recent hackathon, a developer used the tool to clone a voice-to-text application in four hours.
- Cursor is developed by Anysphere, a company founded in 2022 by four MIT graduates: Michael Truell, Sualeh Asif, Arvid Lunnemark, and Aman Sanger. The company reached a $29.3 billion valuation after a $2.3 billion Series D funding round in late 2025, with investors including Thrive Capital, Accel, and new participants like Coatue, NVIDIA, and Google. - The AI editor is a fork of Visual Studio Code, intentionally building on a familiar interface while integrating AI more deeply into the core workflow. This contrasts with tools like GitHub Copilot, which primarily function as extensions within various existing IDEs. - A key differentiator for Cursor is its focus on codebase-wide context, allowing its AI agents to analyze and make changes across multiple files simultaneously. This makes it particularly suited for complex, large-scale projects, whereas competitors like Copilot are often highlighted for speed in single-file, inline code completion. - Beyond the code editor, Cursor's AI agents can be run headlessly via a command-line interface (CLI) and integrated into automation workflows, such as automatically fixing bugs after a failed CI test or updating documentation. The agents also integrate with other tools like Slack and GitHub for tasks like code reviews. - The company has grown to over 250 employees and surpassed $1 billion in annualized revenue as of late 2025. Its enterprise revenue saw a 100x growth in 2025, with adoption by major companies like OpenAI, Spotify, Uber, and Instacart. - The design philosophy behind tools like Cursor is part of a broader shift in Human-Computer Interaction (HCI) toward Human-AI Interaction (HAII). This approach emphasizes augmenting human developers by reducing cognitive load on tasks like remembering syntax, allowing them to focus on higher-level problem-solving and architectural decisions.
