CodePiler Docs
CodePiler docs
Learn how CodePiler imports repositories, filters noisy files, builds prompt-ready output, and helps developers prepare better AI context from real codebases.
What CodePiler does
CodePiler converts a repository into structured prompt context. You import a folder or ZIP, inspect the repository tree, select the files that matter, wrap the output with a system prompt and task, and export in the format your workflow needs.
The application is optimized for repository understanding, debugging prompts, onboarding context, and AI-assisted documentation and refactoring work. Everything runs client-side — no files leave your browser.
Core workflow
The typical flow is: import → search and filter → select files → set a token budget → add a prompt wrapper → export. The file tree and prompt viewer sit side by side so scope stays visible throughout.
This is the main advantage over manual copy-paste: the output stays grounded in the repository structure, the token count is always visible, and the full prompt — system instruction, code context, and task — is assembled in one place before you send it.
What's new
Prompt wrapper: click the Prompt button in the viewer header to open a modal where you can add a system prompt (prepended) and a task (appended). Both sections ship with presets — 12 language-specific system prompts (JavaScript/TypeScript, Python, Go, Rust, Java, and more) and 12 common task templates (modularize, OOP refactor, write tests, security review, and more).
Token budget: select a model context limit from the dropdown next to the token chip — Claude 200K, GPT-4o 128K, Gemini 1M, GPT-3.5 16K — and a color-coded progress bar tracks your usage toward that limit in real time.
Selected files panel: a collapsible panel pinned to the bottom of the file tree lists every selected file with its individual token cost and a one-click remove button, so you can audit your selection without scrolling through the tree.
Separate preview and select: clicking a filename in the tree now previews it in the editor without automatically selecting it. Use the checkbox to include a file in the export. This makes it easy to inspect a file before deciding whether it belongs in the prompt.
Tree controls: Select All and Clear buttons let you include or exclude the entire repository with one click. Expand All and Collapse All buttons manage folder state across the full tree.
Examples
Debugging a multi-file bug
Select the failing route, service layer, and supporting types. Add a debugging task preset, check the token budget fits your model, then export one focused prompt.
Understanding an unfamiliar codebase
Use tree search to gather entrypoints, config, and key modules. Apply a General system prompt preset and export — the model gets full structural context in a single message.