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

# Analyzing media

> Use the vision engine to analyze images and videos for style, composition, color, and motion details.

TalkCut's [vision engine](/core-concepts/vision-engine) can analyze any image or video on your canvas, breaking it down into detailed creative attributes. This analysis helps you understand what makes a reference work and gives the AI agent better context for generation.

## Analyzing images

To analyze an image, ask the AI agent:

* *"Analyze this image"*
* *"What's the style of this photo?"*
* *"Describe the composition and color palette of this image"*

The analysis covers:

| Attribute         | What you get                                          |
| ----------------- | ----------------------------------------------------- |
| **Style**         | Artistic approach, rendering technique, visual genre  |
| **Composition**   | Layout, focal points, visual balance, rule of thirds  |
| **Color palette** | Dominant colors, temperature, contrast, saturation    |
| **Camera**        | Angle, distance, depth of field, lens characteristics |
| **Mood**          | Emotional tone, atmosphere, lighting quality          |

## Analyzing videos

Video analysis includes everything from image analysis, plus motion-specific attributes:

| Attribute     | What you get                                                |
| ------------- | ----------------------------------------------------------- |
| **Motion**    | Camera movement type, subject movement, speed and direction |
| **Pacing**    | Cut frequency, rhythm, temporal flow                        |
| **Lighting**  | How lighting changes across the clip                        |
| **Narrative** | Scene progression, visual storytelling elements             |

To analyze a video:

* *"Analyze this video clip"*
* *"Describe the camera movement and pacing in this video"*

## Practical uses

### Style matching

Upload a reference image with a style you want to replicate. Analyze it, then use the analysis to guide your next generation: *"Generate a new scene in the same style as \[reference]"*.

### Creative learning

Break down professional work to understand what makes it effective. The analysis reveals compositional choices and technical details you might not notice on your own.

### Consistency checks

Analyze multiple generated images to verify they share consistent style attributes. If something drifts, you can spot it early and adjust.

### Building workspace memory

Analysis results contribute to your [workspace memory](/core-concepts/workspace-memory). The more references you analyze, the better the AI understands your visual preferences.

## Credit costs

| Operation        | Credits |
| ---------------- | ------- |
| Analyze an image | 10      |
| Analyze a video  | 10      |

Media analysis is one of the least expensive operations — it's worth analyzing references early and often.