Metadata-Version: 2.3
Name: inprompt
Version: 1.2.0
Summary: Simple CLI to output source code blocks for LLM prompts.
License: MIT
Keywords: prompt,code,formatting
Author: Dominik Roblek
Requires-Python: >=3.9
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Utilities
Classifier: License :: OSI Approved :: MIT License
Requires-Dist: absl-py
Requires-Dist: loguru
Description-Content-Type: text/markdown

# inprompt

A tiny CLI for outputting files or glob patterns as Markdown code blocks. Useful when
you need to copy code into prompts for large language models (e.g., ChatGPT).

## Usage

From the command line, you can pass files and glob patterns to `inprompt`:

```bash
inprompt pyproject.toml '**/*.py' | pbcopy
```

**Note:** It's important to enclose glob patterns (like `'**/*.py'`) in single or double
quotes. This prevents your shell from expanding the pattern before `inprompt` sees it,
ensuring correct file matching, especially for recursive patterns (`**`).

The `| pbcopy` (or equivalent) then pipes the formatted output directly to your
clipboard:

- **On macOS**, `pbcopy` copies STDOUT to the clipboard.
- **On Ubuntu/Linux**, you can use `xclip`. Define aliases for convenience:

  ```bash
  alias pbcopy='xclip -selection clipboard'
  alias pbpaste='xclip -selection clipboard -o'
  ```

  Then you can use the same `inprompt ... | pbcopy` pattern.

Any matched files will be printed to standard output (and thus copied by `pbcopy`) in
the format:

`````markdown
<filename>
````
<file contents>
````
`````

You can then paste those code blocks into an LLM prompt.

## Installation

### Option 1: With pip

```bash
pip install inprompt
```

This installs the `inprompt` command on your system or in your virtual environment.

### Option 2: With [pipx](https://pypa.github.io/pipx/)

```bash
pipx install inprompt
```

This installs `inprompt` globally, isolated from your system's site packages.

## License

[MIT License](LICENSE)

