CodePiler Docs
Examples and FAQ
See practical CodePiler use cases and common questions about filtering, prompt scope, and how to build better repository context for AI tools.
Practical use cases
A common use case is isolating only the modules involved in an error before asking the model to debug the issue. Another is packaging entrypoints, schemas, and config into a prompt for onboarding or architecture explanation.
The main pattern is the same: reduce noise first, then export only the files that support the question you want the AI to answer.
Choosing the right prompt scope
Large prompts are not automatically better. In most sessions, the best result comes from selecting the files closest to the behavior, interface, or subsystem under discussion.
Use tree search and file clicks to validate that the final output contains the exact code path you meant to include.
Examples
Trace an auth flow
Search for auth-related files, select the route, middleware, and service layer, then ask the model for a flow explanation or bug analysis.
Document a feature slice
Select one feature folder, supporting types, and relevant config, then export a structured snapshot for documentation drafting.
FAQ
Should I include every dependency file in the prompt?
Usually no. Include the files that define the behavior, contracts, or data flow you want the model to reason about. Too much unrelated code can dilute the result.
Can I reuse the same repository export across models?
Yes. That is one reason the format switcher exists. You can keep the same selection set and change the export format based on the tool you plan to use.