> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://fastpix.com/docs/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://fastpix.com/docs/_mcp/server.

# AI scene analysis

Scene Analysis is a FastPix In-Video AI feature that parses a video into its individual shots and scene transitions. It returns a chronological breakdown of the video's visual composition, with the start time, end time, and an AI-written description of each distinct scene.

Use Scene Analysis to:

* Build interactive video navigation and chapter overlays based on visual changes.
* Generate a video index so search systems can jump directly to a specific scene.
* Analyze video structure for editing, auditing, and content classification.

> Note: You can generate AI features only for existing media. Upload and process your video first, then generate scene analysis on the ready media.

\


## Before you begin

Make sure you have the following:

* A FastPix account with an active workspace. See [Activate your account](/docs/getting-started/activate-your-account).
* A ready video asset in your workspace.
* To use the API, your **Access Token ID** and **Secret Key**, the media's `mediaId`, and a webhook endpoint if you plan to consume the `video.media.ai.scene_analysis.ready` event. See [get your Access Token ID and Secret Key](/docs/getting-started/activate-your-account#generate-api-credentials) and [Set up webhooks](/docs/webhooks/set-up-webhooks).

\


## How scene analysis works

Scene analysis runs asynchronously on a ready video. You can start it from the FastPix dashboard or with the API. When it finishes, the dashboard shows the scenes on a timeline and as shot descriptions, and the API delivers them through a webhook and a `GET` request.

\


## Generate scene analysis from the dashboard

1. In the FastPix dashboard, open **Video** > **Media** and select the video you want to analyze.

2) In the media's left navigation, under **In-Video AI**, select **Scene Analysis**.

3. Select **Generate**.

The status shows **Processing** while FastPix analyzes the video. When it's done, the results appear on the same page.

### View the results

The results have two tabs:

* **Timeline**: plays the video above a timeline of detected shots (Shot 1, Shot 2, and so on). Hover a shot to see its description, time range, and duration.

- **Shot descriptions**: lists each shot with its time range and AI-written description, for example "Shot 1 · 00:00:00 to 00:00:11".

\


## Generate scene analysis with the API

You can also trigger scene analysis and retrieve the detected scenes programmatically with the In-Video AI API. For the endpoints, request and response schemas, and the `video.media.ai.scene_analysis.ready` webhook, see the API reference: [Generate AI scene analysis](/docs/in-video-ai-api/scene-analysis/create-scene-analysis) and [Get scene analysis](/docs/in-video-ai-api/scene-analysis/get-media-scene-analysis).

\


#### Frequently asked questions

Scene analysis (scene detection) parses a video into its individual shots and scene transitions. For each scene, FastPix returns the start time, end time, and an AI-written description of what's on screen.

No. You can generate scene analysis from the dashboard under **In-Video AI** > **Scene Analysis** for any ready video, and view the shots on a timeline. The API is there when you want to automate the same workflow.

Scene analysis runs asynchronously. In the dashboard, the status moves from **Processing** to the results view. With the API, FastPix sends the `video.media.ai.scene_analysis.ready` webhook when the job completes. See [Get scene analysis](/docs/in-video-ai-api/scene-analysis/get-media-scene-analysis) for details.

Each segment has a `startTime` and `endTime` in seconds and a `description`. The times let you place chapter markers or seek directly to a scene, and the description supports search and classification.

A job is `PENDING`, `PROCESSING`, `COMPLETED`, or `FAILED`. The `result.scenes` list is present only when the status is `COMPLETED`.

Scene analysis detects visual shot changes and describes each one, which is useful for indexing and structural analysis. Chapters group a video into titled, navigable sections for viewers. See [Generate video chapters](/docs/in-video-ai/generate-video-chapters).

\


## What's next

* [Generate video chapters](/docs/in-video-ai/generate-video-chapters) to turn a video into titled, navigable sections.
* [Generate a video summary](/docs/in-video-ai/generate-a-video-summary) for a concise text abstract of a video.
* [In-Video AI overview](/docs/in-video-ai/overview) for the full set of AI features.