Anthropic Unveils Web-Based Claude Code to Challenge GitHub Copilot
Anthropic has unveiled the web-based version of Claude Code—its rapidly growing AI assistant for programmers—now accessible directly through the browser. Beginning Monday, the new platform became available to subscribers of the Pro, Max, and Max+ tiers, priced between $20 and $200 per month. Users can launch the web version via the claude.ai website by selecting the “Code” tab, or through the iOS mobile application.
Previously available only as a CLI tool launched from the terminal, Claude Code has now evolved as Anthropic bets on enabling developers to create and manage their own AI agents within a familiar environment—without the need for additional software installation. This move marks a key step in Anthropic’s broader strategy to expand the Claude ecosystem into a fully fledged, cross-platform development environment.
The market for AI coding tools is growing at an extraordinary pace. GitHub Copilot has lost its monopoly, while Google, OpenAI, Cursor, and now Anthropic have introduced their own solutions. Among them, Claude Code has shown explosive growth over recent months—its user base has increased tenfold since its public launch in May, and the product now generates over $500 million in annual revenue for the company.
Product manager Kat Wu attributed Claude Code’s success to the synergy between Anthropic’s highly precise models and the effortless user experience they deliver. According to Wu, the team’s goal is to make the service not only powerful but genuinely enjoyable to use. Despite the introduction of the web version, she believes the terminal will remain Claude Code’s primary platform, offering developers maximum flexibility and customization potential.
Interestingly, around 90% of Claude Code’s own codebase was written using Anthropic’s models. Wu—formerly an engineer herself—admitted that she has almost entirely stopped writing code manually, instead reviewing AI-generated solutions.
Claude Code and similar tools signal a paradigm shift: while early coding assistants merely completed lines of code, modern AI agents can now autonomously execute complex tasks, effectively transforming engineers into managers of AI systems. Yet not all developers have embraced this transformation. Some studies suggest that productivity may decline when working with AI coders, due to latency in model responses or the need to correct large-scale errors.
Nevertheless, companies continue to invest heavily in this technology. Dario Amodei, Anthropic’s CEO, has previously stated that AI will soon write up to 90% of all code. While this is already nearly a reality within Anthropic itself, the broader industry still has a long road ahead before reaching that future.