A few weeks ago, we opened Jockey’s research preview with a simple idea: working with your videos should start with a conversation. The preview gave us a chance to listen, learn, and refine the experience, including something we heard clearly: you want more precision and control over what Jockey puts together.

Today, we’re bringing Jockey, TwelveLabs’ video intelligence agent, into general availability.

This release brings Jockey to ChatGPT alongside Claude, makes it easier to bring in and explore your videos, and gives you more control over the clips you turn into a finished reel. We’ve also added ways to share both the videos you create and the insights you find.

Here are seven updates we’re shipping.

1. Jockey is now available in ChatGPT, in addition to Claude

Jockey is now available in ChatGPT through its MCP connection, in addition to Claude. Connect your TwelveLabs account, choose a video collection, and start asking questions.

In ChatGPT: You can go to Settings → Plugins → Browse Directory → Search for ‘Jockey by TwelveLabs’

In Claude: You can go to Customize → Connectors → Discover → Search for ‘Jockey by TwelveLabs’

2. Import videos directly from social URLs, Google Drive, or your device within the MCP

Start with videos on YouTube, in Google Drive, or saved on your device.

• For online videos, paste URLs into ChatGPT or Claude and say: “Jockey, add these videos to my collection.” You can use a YouTube link or a direct link to a video file.

• For videos in Google Drive or on your device, ask Jockey for an upload link, then choose the files you want to add.

Once your videos finish processing, you’re ready to ask questions, find specific moments, and create clips.

3. View scene-by-scene breakdowns directly in your conversations

Ask Jockey to break down a video, and explore the result as an interactive visual answer inside ChatGPT or Claude. Scene previews, descriptions, and video references now appear directly in your conversation.

• Try: “Break this video into scenes and describe what happens in each one.”

You can scan the breakdown, identify a scene, and inspect the relevant footage. This makes it easier to understand how a video is structured or choose moments for a new edit.

4. Clip with sub-second precision

A fraction of a second can change how a clip feels. When you turn your selected moments into a reel, Jockey preserves their precise start and end points.

A clip selected from 12.4 to 16.8 seconds keeps those boundaries, helping you retain the end of a sentence or catch the moment an action begins.

5. Shape your video with timeline editing

Users wanted more control over what Jockey puts together. That’s why we added timeline editing, so you can bring your own judgment and precision to the final cut.

Once Jockey assembles your reel, trim the beginning or end of a clip, drag clips into a new order, and preview how the sequence plays.

Put your strongest moment first, shorten sections that run too long, and arrange the footage to tell the story you want.

6. Download your finished reel as an MP4

When your reel is ready, download it as an MP4 directly from ChatGPT or Claude. Use the finished video in a presentation, campaign, or social post.

7. Share Jockey’s responses with a link

Save your analysis as a shareable Jockey artifact and send it to your team. They can reopen the results, explore your findings, and review the video moments behind them, making it easier to discuss insights and decide what to use next.

Put Jockey to work on a video you already have. Connect it in ChatGPT or Claude, add a product demo, interview, or campaign video, and ask: “Find three moments worth sharing and show me the clips.” Review the results, build a reel from your selections, and refine it before exporting an MP4 or sharing a Jockey artifact with your team.

Get started with Jockey.

Need help connecting? Follow the setup guide for ChatGPT and Claude.

Watch Jockey in action: