Jockey by TwelveLabs - Your entire video library: organized, searchable, understood
by•
Jockey is the first AI that understands your entire media library across photos and video. It analyzes thousands of hours of footage so you can find and connect people, moments and context across videos. Use it in Claude or ChatGPT, or build with the API/SDK. Ingest YouTube, TikTok and Instagram Reels links; assemble content into MP4s; and create and share video-driven artifacts with playable clips. Every TwelveLabs model upgrade makes Jockey better, without re-integration.
Replies
We're launching the general availability of Jockey today, and it's shaped by what we learned during the research preview. We heard three things loud and clear:
People needed the connectors to work better. The strongest reactions came through Claude and ChatGPT. So we prioritized ChatGPT support for GA and made the MCP experience seamless. If you want to query your library, your chat interface should be the last friction point, not the first.
Builders needed output they could control. Developers told us a black-box "final edit" wasn't what they wanted. They needed structured clips, JSON, artifacts they could manipulate. So we built durable artifacts that can be saved, shared, revoked, or deleted and ensured the API gives you timestamped, actionable output you own, not a render you're stuck with.
Users wanted deep social intelligence, not just basic video search. Creators and marketers kept asking us: "Why is this video trending? What specific hooks are holding attention? What common patterns make these reels go viral?" We listened, and built social intelligence features that analyze video hooks, breakdown visual and audio pacing, and explain what actually works across trending content. We also added social ingestion of URL links from YouTube, TikTok, and Instagram Reels, so users can analyze what's working in the wild and search performance alongside content without having to upload videos. It's a whole new way to deconstruct viral video performance.
Library-level understanding is the point. You point Jockey at thousands of videos and images, and it reasons across all of them, not just one. That's why we built it this way, and why every model upgrade makes it better without you lifting a finger.
If you're building a video app, organizing footage, or trying to find something in thousands of hours of content, Jockey is for you. Try Jockey and let us know you feedback and we will fix it!
S/O for this new launch! keep up the great work, and keep launching
thanks for the continuous support!
@hamza_afzal_butt Great question! Long videos are what we built Jockey for.
Ask about a specific moment and Jockey finds it, even in a video that runs for hours (~10GB per video) and in a library spanning thousands hours of videos. Every answer cites the clips it used with start and end times, so you can jump straight to the moment and check it.
Speaking of search specificity, it works best when the moment has something concrete to find, like a line of dialogue, an action, or an object on screen.
This new launch by @TwelveLabs connects the dots.
Marengo (3.0) launched 10 months ago, ranked among the Top Products of the Day. Pegasus launched 6 months ago (1.5) and 1 week ago (1.6), and again, both ranked among the Top Products of the Day.
S/O to ?makers for their consistency. It pays off. Keep launching.
Congrats on the launch! I've always wanted videos to be searchable! Although currently my footage as scattered all over the place. Time to get organized.
@heyitsirenechan time to get organized, exactly! try for free here: playground.twelvelabs.io/get-started
Congrats on the launch! Being able to find specific people and moments across thousands of hours of footage and turn them into shareable clips sounds really useful.
thanks for the support! try it for yourself on the playground at twelvelabs.io
@tessak22 thanks for your continuous support, Tessa! you're the best
What is the workflow of a standard Claude or ChatGPT, and how big a library does it take before searches become slow?
Additionally, Congratulations @jaelee_ , @aiden_lee7 and team Jockey by TwelveLabs 🚀✌️
@aymi_malik excellent questions. Typical workflow starts with
creating a knowledge store and adding media
asking jockey for what you want from that knowledge store
saving the output as an artifact or downloading the mp4 file that stitches together the cited clips.
Please check out
if you have not!For your second question: "how big a library does it take before searches become slow", the search and query latency does not grow with the library's size. However the query may take longer depending on the complexity of the ask that requires more reasoning steps and tool calls along the route.
Let me know if you have any other questions!
Video is the one area that I feel AI hasn't yet reached its potential in and this tool seems to go in that direction.
Congrats on the launch team!
@sumitsaurabh927 strong +1 and @TwelveLabs is a pioneer here. follow for future launches!
@jaelee_ @aiden_lee7 How well does TwelveLabs understand long videos and find specific movement?
oh wow, searchable videos? I have a ton of use cases for these. Congrats on the launch.