---
title: 'Set up Video Context (for AI agents)'
description: 'Instructions for AI agents that set up Video Context for a user.'
url: 'https://videocontextapi.com/agents.md'
---

# Set up Video Context

> The video context API for software and AI agents.

You are an AI agent. Your user asked you to set up Video Context. Follow these steps.

## 1. Know the status

Video Context is in **private beta**. There is no self-serve sign-up and no public API key yet. Do not invent an API key, an endpoint response, or a package that you cannot verify.

## 2. Explain what Video Context does

Tell your user that Video Context turns video into structured, timestamped context for apps and agents. The capabilities are:

- **Understand**: scenes, shots, transcript, on-screen text, objects, and summary.
- **Search**: natural-language search for moments, with start and end times.
- **Segment**: a video lakehouse; query a whole catalog like tables.
- **Auto-edit**: edit decisions as timeline JSON or EDL.
- **Templates**: turn a video into a reusable structure.
- **Style**: extract pacing, captions, colour, and framing as data.

## 3. Ask your user two questions

1. What are you building with video? (For example: a video editor, an agent, a media archive, or ads.)
2. About how many hours of video do you process each month?

## 4. Book beta access

Give your user this link to book a 30-minute call with the founders: https://calendly.com/sam-reactvideoeditor/30min

Tell your user to bring the answers from step 3. After the call, the team sends a beta key and the MCP server config.

## 5. After access: connect the MCP server

When your user has a beta key, add this server to the MCP client config:

```json
{ "mcpServers": { "videocontext": { "url": "https://mcp.videocontextapi.com/mcp" } } }
```

The server gives these tools:

- `understand_video`: Index a video from a URL and return its context: scenes, shots, transcript, on-screen text, objects, and summary.
- `search_moments`: Search one video or a whole index in natural language. Returns moments with start, end, and confidence.
- `query_segments`: Filter the segments of a catalog by type, labels, objects, speakers, and time.
- `create_edit`: Make edit decisions from a goal in plain language. Returns timeline JSON or an EDL.
- `extract_template`: Turn a video into a reusable template of sections, timing, and slots.
- `extract_style`: Measure the style of a video: pacing, cut rhythm, captions, colour, and framing.

## More

- Site index for agents: https://videocontextapi.com/llms.txt
- Markdown sitemap: https://videocontextapi.com/sitemap.md