- COLLECTIONS
- INDEXED
- IDENTIFIED
- HUMAN
Short answer
How can I track people or objects in my videos using AI?
Your AI uses Twelve Labs to index your video files and group specific subjects into organized collections. You'll be able to find every instance of a person or object without watching every frame manually. Here is how the process works.
Where every finding lands.
The outcomes of entity tracking
The AI organizes your video data into these specific formats.
COLLECTIONS
Grouped entity sets
The AI builds sets of specific subjects. You get a structured way to manage all clips featuring a single person or object.
INDEXED
Searchable video assets
Your video files become searchable data. The AI processes the footage so you can query it later.
IDENTIFIED
Specific entity profiles
Individual subjects are defined as unique entities. This allows the AI to distinguish between different people or items.
HUMAN
Manual review needed
You decide which entities are important enough to save. The AI flags the subjects, but you control the final organization.
The workflow
What your AI does when <the thing arrives>.
The AI handles the heavy lifting of visual recognition and data organization.
Index the footage
The AI scans your video files to understand the visual content and prepare it for searching.
get_indexed_assetDefine a subject
Once a subject is found, the AI creates a unique identity for that person or object.
create_entityBuild a group
The AI gathers all relevant instances of a subject into a single, manageable set.
create_entity_collectionVerify the data
The AI presents the identified entities and collections for you to check and refine.
get_indexed_asset
Try it
Copy these to start.
Use these prompts to trigger the tracking process.
Starting points
These are basic starting lines. Swap out the bracketed text with your actual video folder names or subject descriptions.
Twelve Labs Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_Vgl5VVTXIogBXlRhR1Xxpl2YsHWFIR0xQ7beKVtS/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Twelve Labs capabilities are ready to use.
{
"mcpServers": {
"twelve-labs-video-understanding-mcp": {
"url": "https://edge.vinkius.com/vk_preview_Vgl5VVTXIogBXlRhR1Xxpl2YsHWFIR0xQ7beKVtS/mcp"
}
}
}Find all clips in the 'Security_Footage' folder that contain a person wearing a red hat.
Create a new entity collection for all instances of the 'Delivery Truck' in my recent uploads.
Identify the specific person appearing in the 'Interview_Series' assets.
Look through the 'Event_Video' files and create an entity for the keynote speaker.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Twelve Labs once, then ask.
Just link your account once. Your credentials stay encrypted, and you can immediately start asking your AI to find things in your videos.
Connect Twelve Labs to your AIFAQ
How this task behaves
- 01
Can the AI delete my original video files?
No. The AI only reads and indexes the files to create metadata; it cannot modify or delete your source footage.
- 02
Does the AI decide which entities are important?
No. The AI identifies potential subjects, but you must tell it which ones to turn into entities or collections.
- 03
Can I track objects that aren't in the video?
No. The AI can only track and index what is physically present in the video files you provide.
- 04
Will the AI automatically organize my entire hard drive?
No. You must point the AI toward specific assets or folders to begin the indexing process.
- 05
Can the AI rename my video files?
No. The tools are designed for indexing and entity management, not for file system manipulation.
More questions about Twelve Labs? The Connector page answers them. See everything the Twelve Labs Connector can do
Connect Twelve Labs to Claude, Cursor, ChatGPT & more
