Bolt Integrates Claude Sonnet 4 With 1M Context Window for Full Codebase Intelligence.

Bolt has integrated Claude Sonnet 4's expanded context window into its AI-powered development platform, giving users access to what the company describes as the most comprehensive codebase intelligence available in any building tool.

The integration centers on Claude Sonnet 4's newly expanded 1 million token context window, which allows Bolt's AI agent to ingest and analyze entire codebases in a single pass. The company announced the update on LinkedIn, highlighting the agent's ability to understand project architecture and dependencies across thousands of files.

For developers working on multi-component applications, this means the AI can now trace how changes in one part of a codebase affect others, maintain consistent patterns across features, and avoid breaking existing functionality when generating new code. The expanded context effectively eliminates the need to manually guide the AI through complex project structures.

Previous versions of Bolt's AI operated with smaller context windows, requiring users to provide explicit instructions about file relationships and architectural decisions. With the 1M token window, the agent can now read documentation, configuration files, component libraries, API routes, and database schemas simultaneously.

This matters most for teams building production applications rather than quick prototypes. A typical React application with authentication, database integration, and multiple feature modules can easily span hundreds of files. Being able to reason about the entire structure lets the AI make informed decisions about where to place new code, which existing patterns to follow, and how to maintain type safety across modules.

The feature builds on Bolt's existing model integrations, which already include multiple Claude variants and OpenAI models. Users working on bolt.new can now select Sonnet 4 as their primary model, with the platform automatically leveraging the full context window when analyzing projects.

The update arrives as Bolt continues expanding beyond rapid prototyping into production-ready development. The platform recently launched cloud hosting integrations with Netlify and Supabase, addressing earlier criticism that projects built in Bolt often struggled to escape "prototype purgatory."

Developers using competing AI coding tools like Cursor, Windsurf, or Lovable now have another reason to evaluate Bolt's approach. While those platforms offer different workflows and feature sets, the combination of full codebase context and Bolt's prompt-to-production model creates a distinct building experience.

The integration is live now for all Bolt users at bolt.new, with no additional configuration required beyond selecting Claude Sonnet 4 as the active model in project settings.