TLDR: Cursor, developed by Anysphere, has launched a new web application that allows developers to manage AI coding agents directly from any web browser, including mobile devices. This expansion beyond its traditional IDE offers enhanced flexibility, enabling users to assign tasks, monitor progress, and integrate code changes remotely. The move follows significant financial milestones and aims to make AI-assisted development more accessible and collaborative.
San Francisco, CA – July 6, 2025 – Anysphere Inc., the innovative company behind the popular AI-powered coding editor Cursor, has officially launched a new web application designed to empower developers with seamless management of AI coding agents. This strategic expansion marks a significant evolution for Cursor, moving beyond its core Integrated Development Environment (IDE) to offer unparalleled flexibility and accessibility for AI-assisted software development.
The new web app allows users to assign complex programming tasks to AI agents using simple natural language commands. Whether working from a desktop, tablet, or mobile device, developers can instruct agents to write new features, fix bugs, answer codebase questions, or scaffold project components. A key feature of the platform is the ability to monitor agent progress in real-time and directly merge completed code changes into projects, streamlining the development workflow.
This release builds upon Cursor’s recent advancements, including the introduction of ‘background agents’ in May 2025, which autonomously handle coding assignments, and a Slack integration launched in June 2025, enabling task assignment by tagging @Cursor, similar to other advanced AI agents like Devin. The web application centralizes these capabilities, providing a unified dashboard for tracking and collaborating on agent-driven tasks. Each agent’s run can be shared via a unique link, facilitating team collaboration and code review processes.
Anysphere has seen remarkable growth, recently closing a substantial $900 million funding round that valued the company at an impressive $9.9 billion. The company has also surpassed $500 million in annualized revenue, with over half of Fortune 500 companies reportedly utilizing Cursor’s solutions. Major tech firms such as Nvidia, Uber, and Adobe are among its growing client base, underscoring the widespread adoption of its AI development tools.
According to Michael Truell, CEO of Cursor, AI agents are poised to revolutionize the industry. “AI agents could handle up to 20% of programmers’ work in the near future,” Truell stated, highlighting the potential for increased efficiency and faster workflows. Andrew Milich, Anysphere’s Head of Product Engineering, emphasized the user-centric motivation behind the web app, explaining that customers desired Cursor to operate in more environments and solve a broader range of problems, thereby “removing the friction” from their development processes.
While the core Cursor IDE remains a powerful tool, the web app signifies Anysphere’s commitment to making AI coding agents accessible from anywhere. The web platform is built as a Progressive Web Application (PWA), offering a native-like experience on iOS and Android, complete with push notifications for task completion and offline viewing capabilities for past agent runs. This ensures developers can stay connected to their projects even when away from their primary development environment.
Also Read:
- Google Unveils Gemini CLI: An Open-Source AI Powerhouse for Command Line Developers
- Microsoft Unveils Open-Source GitHub Copilot Chat for Enhanced AI Transparency in VS Code
Access to the agent management features within the new web app is available to paid subscribers. Cursor offers a $20 monthly Pro tier, with a premium $200/month Ultra tier now available for power users and larger enterprises, providing advanced functionality and greater scale. Free accounts do not include access to these agent management capabilities. Anysphere’s leadership, including CEO Ben Thompson, has stressed the importance of waiting for AI reasoning models to mature sufficiently to ensure the accuracy and reliability of their agents, a milestone they believe has now been achieved, promising practical and dependable tools for the developer community.
