- What is Context Bundle?
- Context Bundle is a free, browser-based tool that packages a codebase (or just a few files) into a single well-formatted prompt you can paste into any large language model. It runs entirely in your browser — no code is uploaded to any server.
- How are tokens estimated?
- Files are counted with a real byte-pair encoder (via gpt-tokenizer), not a naive character heuristic. The result is presented as a size bucket (Tiny, Small, Medium, Large, Huge) so it is meaningful across any modern LLM.
- Does Context Bundle send my code to a server?
- No. All file reading, tokenization, and bundle generation happens client-side in the browser. There is no upload step and no analytics on your source code.
- What files are included by default?
- Common text and source-code file types (JavaScript, TypeScript, Python, Go, Rust, Markdown, JSON, HTML, CSS, shell, etc.). Common junk directories — node_modules, .git, dist, build, .next, venv — are skipped automatically. You can add additional glob patterns like *.log or **/*.test.ts.
- Can I pick individual files from a dropped folder?
- Yes. Drop a folder, expand the card, and toggle any file on or off. The token count and size update live.
- Which LLMs does the bundle work with?
- Any model that accepts a text prompt: GPT, Claude, Gemini, Llama, Mistral, local models, and more. The output is plain Markdown with file paths and fenced code blocks.
- Is Context Bundle free and open source?
- Yes. The project is MIT-licensed and the source is available on GitHub.