Anthropic has launched a web-based version of its popular AI coding assistant, Claude Code, allowing developers to create and manage multiple AI coding agents directly from their browser. The new web app is now being rolled out to subscribers of Anthropic’s $20-per-month Pro plan, as well as its higher-tier $100 and $200-per-month Max plans. Users can access the feature by visiting claude.ai and selecting the “Code” tab, or through the Claude iOS app. This move marks a significant step in Anthropic’s effort to expand Claude Code beyond its original command-line interface (CLI) roots. By bringing the tool to the web, the company aims to make AI coding agents more accessible and easier to use across different development environments. The shift comes amid growing competition in the AI coding space, where tools from Microsoft’s GitHub Copilot, Cursor, Google, and OpenAI have already established strong footholds—many of them available on web platforms. Despite the crowded market, Claude Code has emerged as one of the most widely adopted AI coding tools. Since its broader launch in May, user numbers have grown tenfold, and the product now generates more than $500 million in annualized revenue for Anthropic. According to Cat Wu, Anthropic’s Product Manager, a key reason for the tool’s success lies in the quality of the company’s AI models, which developers have increasingly favored. But Wu also emphasized that the team has intentionally designed the product to be engaging and enjoyable. “We try to sprinkle in some fun wherever we can,” she said, highlighting a user experience that goes beyond pure functionality. While the web and mobile versions are a major focus, Wu stressed that the command-line interface will remain the core experience for advanced users. “As we look forward, one of our key focuses is making sure the CLI product is the most intelligent and customizable way to use coding agents,” she said. “But we’re continuing to put Claude Code everywhere, helping it meet developers wherever they are.” Anthropic claims that 90% of the Claude Code product itself was written by its own AI models. Wu, a former software engineer, said she now rarely types code herself, instead spending most of her time reviewing and refining the AI’s output. Early AI coding tools functioned like smart autocompleters, suggesting code as developers typed. In contrast, agentic tools like Claude Code allow developers to deploy autonomous AI agents that can plan, write, test, and debug code independently. This shift has transformed the role of many software engineers into that of AI managers, overseeing multiple agents rather than writing code from scratch. However, the transition hasn’t been seamless for everyone. A recent study found that some engineers were actually slower when using AI coding tools like Cursor, partly due to time spent prompting and waiting for responses. Additionally, AI agents often struggle with large, complex codebases, leading to errors that require significant human oversight. Despite these challenges, companies like Anthropic remain committed to advancing AI coding agents. Anthropic CEO Dario Amodei recently predicted that AI could eventually write 90% of code for software engineers. While this may be true within Anthropic’s own development workflows, widespread adoption across the broader tech industry is likely to take more time.