> ## Documentation Index
> Fetch the complete documentation index at: https://docs.wearer.xyz/llms.txt
> Use this file to discover all available pages before exploring further.

# Generate Guide

> Generate a step-by-step guide from processed video data

# Generate Guide Command

The `generate-guide` command creates a detailed step-by-step guide from previously processed video data. It uses AI to analyze the extracted frames and generate a structured guide.

## Usage

```bash theme={null}
transfer-learning generate-guide PATH [OPTIONS]
```

## Arguments

| Argument | Description                            |
| -------- | -------------------------------------- |
| `PATH`   | Path to processed video data directory |

## Options

| Option                  | Description                                  | Default     |
| ----------------------- | -------------------------------------------- | ----------- |
| `--output-dir TEXT`     | Output directory for guides                  | `guides/`   |
| `--model TEXT`          | Model to use for guide generation            | From config |
| `--temperature FLOAT`   | Temperature for generation                   | 0.2         |
| `--user-directive TEXT` | User directive to customize guide generation | None        |
| `--skip-cache`          | Skip cache and force regeneration            | False       |
| `--format TEXT`         | Output format (json, markdown, html)         | json        |
| `--help`                | Show help message and exit                   | -           |

## Examples

### Generate a guide from processed video data

```bash theme={null}
transfer-learning generate-guide data/my_video
```

### Generate a guide with a custom directive

```bash theme={null}
transfer-learning generate-guide data/my_video --user-directive "Focus on technical details and tools used"
```

### Generate a guide in markdown format

```bash theme={null}
transfer-learning generate-guide data/my_video --format markdown
```

### Generate a guide with a specific model

```bash theme={null}
transfer-learning generate-guide data/my_video --model gpt-4-vision-preview
```

## Output

The command generates a guide file in the specified format (default: JSON) in the output directory:

```
output_dir/
└── video_name_guide.json
```

## Guide Format

The generated guide has the following structure:

```json theme={null}
{
  "title": "Process Title",
  "description": "Overall process description",
  "steps": [
    {
      "step_number": 1,
      "title": "Step Title",
      "description": "Step description",
      "tools_used": ["tool1", "tool2"],
      "duration": 10.5,
      "key_points": ["point1", "point2"],
      "timestamp": 0.0
    }
  ],
  "principles": [
    {
      "name": "Principle Name",
      "description": "Principle description",
      "importance": "Why this principle matters",
      "examples": ["example1", "example2"]
    }
  ],
  "total_duration": 120.5,
  "source_type": "local",
  "source_path": "path/to/video.mp4"
}
```

## Customizing Guide Generation

You can customize the guide generation process using the `--user-directive` option. This allows you to specify what aspects of the video to focus on or what style to use for the guide.

Examples of user directives:

* "Focus on technical details and tools used"
* "Emphasize beginner-friendly explanations"
* "Highlight safety considerations"
* "Include more detailed steps for complex parts"

## Next Steps

After generating a guide, you can:

1. Use the guide in your application or documentation
2. Convert the guide to different formats using the `--format` option
3. Customize the guide further with manual edits

## Related Commands

* [Process Video](/cli/process-video)
* [YouTube Guide](/cli/youtube-guide)
* [Analyze](/cli/analyze)