Anthropic Expands AI Coding Assistant "Claude Code" to Web and iOS
Anthropic has announced the expansion of its AI-powered coding assistant, Claude Code, to iOS and web platforms. Previously, the tool was limited to terminal and desktop development environments. This update allows users to connect to code repositories via natural language, issue coding commands, and let the AI automatically write, edit, and test code—all through a browser or mobile device. The new version brings Claude Code’s core capabilities into a flexible, cloud-based workspace accessible on any platform, whereas previously it was offered only as command-line and IDE integrations for developers.
With the new web version, Claude Code now offers a complete coding interface similar to a real-time development environment. Users can link their GitHub projects, instruct the model to perform tasks such as fixing bugs, refactoring code segments, adding new features, or optimizing performance, and then watch as Claude Code automatically implements the changes. The AI can run tests, create commits, and open pull requests without manual input, operating within a secure cloud sandbox that simulates a real coding environment. This approach eliminates the need for local setups and complex integrations.
The web version's interface features multiple panes, allowing users to view prompts, AI responses, and code outputs side by side. Each coding session runs in an isolated environment, with configurable network access options based on the project’s security needs. Early previews show that users can choose between fully sandboxed sessions with no internet access or open settings allowing limited external access. The release is currently available via an early access program, but the company has confirmed a full public rollout is coming soon.
On the iOS platform, Anthropic has integrated Claude Code directly into the Claude app. While not designed to replace full desktop IDEs, the iOS version enables users to review code changes, inspect pull requests, and trigger small fixes from anywhere. Developers can start a task on a laptop and later continue managing it on their phone. This mobility is especially useful for distributed teams and remote work environments.
This development comes amid increasing competition in the AI coding tools space, with Anthropic positioning Claude Code as a strong contender against GitHub Copilot and OpenAI’s Codex. The timing of this release is particularly significant as the company recently launched its latest AI model, Claude Sonnet 4.5, while also facing several controversies and challenges. Financially, despite revenue growth from around $1 billion in early 2025 to over $5 billion by August, the company remains unprofitable due to high costs related to computing power, data processing, and hiring top AI researchers. Even after settling an author copyright lawsuit for $1.5 billion, Anthropic continues to face major legal hurdles, including a case filed by Reddit in June 2025 alleging unauthorized use of content.