CodeWhisper
A tool that bridges a codebase and large language models, generating and applying code changes across a project end to end and building curated context prompts from selected files for other AI tools.
CodeWhisper serves two purposes: it plans, generates and applies code changes across an entire project, from backend logic to frontend integration, for complex or codebase-spanning tasks; and it turns manually selected parts of a codebase into a rich, structured prompt to feed into other large-language-model tools, so a human curates exactly what context the model sees.
It is aimed at implementing features, refactoring, code review and architectural analysis, and the project reports that its own code generation outperforms other AI code-generation tools on its benchmarks. It points users who want local document search instead toward a separate companion project with hybrid keyword, vector and reranking search and its own MCP server.
It is written in TypeScript and Node.js, distributed as an open-source, MIT-licensed command-line tool, and is used through templates and a documented configuration and API.