AI Prompt for Writing Docstrings

Generate professional docstrings with this AI prompt: Google, NumPy or JSDoc style, typed params, edge cases, raises sections, and runnable examples.

The prompt

Role: You are a documentation perfectionist who writes docstrings that make APIs a joy to use.

Context: Language: [LANGUAGE]. Convention: [CONVENTION: Google / NumPy / JSDoc / Javadoc / etc.].

Code:
[CODE: function/class/module]

Task: Write the complete docstring including: 1) One-line summary + longer description of what it does and when to use it. 2) Every parameter: type, meaning, constraints, default behavior. 3) Return value: type and meaning. 4) Exceptions/errors raised and when. 5) Edge-case behavior. 6) A runnable usage example. 7) Any "see also" references to related functions.

Format: The docstring in a code block, ready to paste above the code, followed by a note on anything you inferred (vs explicitly in the code).

Ask the AI to mark every inferred type and edge case with a confidence note, so you can verify the guesses that matter. Docstrings built on wrong inferences mislead every future reader.

Constraints: Follow [CONVENTION] exactly — this feeds doc generators. Infer types from the code but LABEL inferences so I can correct them. The example must actually run.

Variables to fill in

Replace each [variable] in the prompt with your own details:

[LANGUAGE] [CONVENTION] [CODE]

Example input

LANGUAGE=Python, STYLE=Google docstrings, CODE=def train_model(data, epochs=10, lr=0.001) with early stopping, INCLUDE=Args, Returns, Raises, Example

How to use this prompt

Name your docstring convention (Google, NumPy, JSDoc, Javadoc) — format matters for doc generators. List the sections you want; Examples are the most valuable.