What file types and library sizes does Jockey support?

Jockey supports standard video formats (MP4, MOV, AVI, MKV, WebM) and photo formats (JPEG, PNG, HEIC, HEIF, WebP, BMP, GIF). Specific storage and per-file size limits are covered in the documentation.

Does Jockey use facial recognition to identify people?

No. Jockey uses multimodal AI — analyzing visual, motion, and contextual cues — to group people across your library rather than traditional facial recognition. TwelveLabs states it does not build biometric databases, compare data against external records, or share identity information between accounts.

Can developers build Jockey's capabilities into their own products?

Yes. TwelveLabs offers an API and SDK exposing multimodal embeddings (one embedding across video, image, audio, and text), natural-language video search, structured data extraction with timestamped JSON metadata, and frame-accurate timestamps for deep-linking into playback.

Jockey — AI video understanding that works in Claude and ChatGPT

Jockey is TwelveLabs' AI video understanding and marketing intelligence product. It indexes video and photo libraries so you can search moments by meaning (phrase, image, clip, person, place, action), understand what happens inside media scene by scene, automatically organize footage into groups, and assemble finished clips or highlight reels that can be trimmed and exported as MP4. It works inside ChatGPT and Claude via an MCP integration (context carries across sessions), through a developer API/SDK, and adds trend intelligence across YouTube, TikTok, and Instagram Reels for marketing teams. The product is built on TwelveLabs' underlying models: Marengo (indexes media by meaning), Pegasus (understands media in natural language), and Jockey (orchestrates the workflow).

Key Features

Semantic media search

Search across every video and photo in your library by meaning rather than tags or filenames — by phrase, image, clip, person, place, or action.

Deep media understanding

Jockey reads each clip and explains what's there scene by scene and moment by moment, including scenes, context, captions, topics, sentiment, and key moments.

Automatic organization

Footage sorts itself into meaningful sets by person, place, topic, event, character, or project — no manual foldering or naming conventions.

Assembly and export

Ask Jockey to assemble finished media from source footage; refine, trim, and export results as an MP4. Can generate instant highlight reels or edit-ready cuts on a topic.

Works in ChatGPT and Claude

Talk to your footage from the assistants you already use via MCP integration, with context that carries across sessions; also connectable via the TwelveLabs API.

Marketing ad intelligence

Automatic ad tagging by hook, talent, format, and tone; winning hook detection for opening three seconds; talent and logo recognition; format and trend breakdowns; performance insights with transcripts, creators, views, and engagement.

Social and trend intelligence

See what's rising on YouTube, TikTok, and Instagram Reels for your brief, drawn from a social corpus TwelveLabs keeps current.

Concept video previews

Turn a campaign direction into a playable 15–30 second concept video, then save, share, or revoke it.

Pros & Cons

Pros

Works directly inside ChatGPT and Claude, so teams can query video from assistants they already use, with context that carries across sessions
Search is meaning-based (phrase, image, clip, person, place, action) rather than relying on tags or filenames
End-to-end pipeline from search and understanding to organization, assembly, trimming, and MP4 export
Marketing-specific features like ad tagging, hook detection, talent/logo recognition, and performance insights with transcripts and engagement data
Solid developer surface: API, SDK, multimodal embeddings, structured JSON metadata extraction, and frame-accurate timestamps

Who It's For

Best for

Content creators and editors who need to search, quote, and cut from large video archives inside ChatGPT or Claude
Marketing teams that want every ad in their creative library tagged, analyzed, and ranked by performance
Social teams tracking rising formats and trends on YouTube, TikTok, and Instagram Reels
Developers embedding multimodal video search, understanding, and metadata extraction into their own products via API

Not ideal for

Teams that need SSO/SAML or a dedicated environment on lower-tier budgets (those are Enterprise-only)
Users who need predictable overage billing on Free or Plus plans (overage isn't included until Pro)

Use Cases

Finding exact clips, quotes, or cuts in a large archive by describing the moment in plain language

Searching everything said on camera and jumping straight to the line

Generating instant highlight reels from a topic or set of best moments

Tagging ad libraries by hook, talent, format, and tone automatically

Detecting which opening three seconds hold attention across an ad library

Locating every ad a creator, spokesperson, or brand logo appears in

Pricing

Every plan includes storage, media processing, and access to TwelveLabs APIs, priced by media volume and monthly usage. Monthly billing or annual billing at a 17% discount. Free: $0/month with 3 GB cumulative storage, basic workload and rate limits, shared environment. Plus: $29/month ($290/yr) with 50 GB storage, medium workload, high rate limits. Pro: $99/month ($990/yr) with 200 GB storage, hig

Free

$0/month

Explore and test the APIs. Includes 3 GB cumulative storage, basic workload allowance and rate limits, shared environment, no overage, no SSO/SAML.

Plus (Recommended)

$29/month ($290/yr annual)

Build with more storage and usage. Includes 50 GB cumulative storage, medium workload allowance, high rate limits, shared environment, no overage, no SSO/SAML.

Pro

$99/month ($990/yr annual)

Scale with growing usage. Includes 200 GB cumulative storage, high workload allowance, high rate limits, shared environment, optional overage, no SSO/SAML.

Enterprise

Custom

For organizations with custom needs. Custom cumulative storage, custom workloads and rate limits, dedicated environment, overage by agreement, SSO/SAML included.

FAQ

What does Jockey do?

Jockey is TwelveLabs' AI video understanding product. It searches your video and photo library by meaning, explains what happens inside clips scene by scene, automatically organizes footage into meaningful sets, and assembles finished media (clips, collections, timelines, highlight reels) that you can trim and export as MP4. It can be used inside ChatGPT and Claude, via the API, and includes marketing intelligence features like ad tagging and trend tracking across YouTube, TikTok, and Instagram Reels.

How do I get started with Jockey?

Connect via Claude, ChatGPT, or the API (you'll need an API key), then upload videos or paste a public link, and ask Jockey to get your answer. TwelveLabs provides a step-by-step getting-started guide.

How much does Jockey cost?

Plans are Free ($0/month, 3 GB storage), Plus ($29/month or $290/yr, 50 GB), Pro ($99/month or $990/yr, 200 GB), and Enterprise (custom pricing). Annual billing saves 17%. Every plan includes storage, media processing, and access to TwelveLabs APIs.

Can I delete my data, and what gets deleted?

Jockey