Twelve Labs
Video intelligence API that indexes, searches, and analyzes footage at scale with the Marengo and Pegasus models
Twelve Labs remains our pick for enterprise-scale video search, segmentation, and compliance review — the vendor publishes a +13.1% Pegasus 1.5 advantage over Gemini 3.1 Pro on multimodal prompting and #1 on Video-MME, and the any-to-any retrieval across video, audio, image, and text is genuinely differentiated. The math still bites: at $2.50/hour of indexing and $1.75/hour of Analyze input video, a 500-hour archive costs real money before you've run a single query. If you're under roughly 100 hours of footage, a general-purpose video model will be cheaper. For archives in the thousands of hours, few alternatives match the composite accuracy.
Verified 4d ago · liveness 66/100 · cite: rightaichoice.com/tools/twelve-labs
- Media companies and archives with thousands of hours of footage
- Adtech and contextual targeting platforms
- Government and public sector evidence teams
- Developers building video search or analysis products
- Hobbyists with under 100 hours of video
- Teams wanting a no-code visual interface rather than an API
- Organizations with minimal video volume where pay-as-you-go indexing outweighs value
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Twelve Labs if your library is under roughly 100 hours and you want a no-code interface — pay-as-you-go indexing at $2.50 per video hour plus $1.75/hour Analyze input and $4 per 1,000 searches will exceed what a general-purpose video model costs you.
Video indexing is a one-time $2.50 per hour of video on the Developer tier, so a 500-hour archive costs $1,250 before you run a single query.
The Free tier (10 hours of indexing, 90-day index access) fits evaluation and pilot work. The Developer pay-as-you-go tier — $2.50 per hour of video indexed, $1.75/hour Analyze input video, $4 per 1,000 queries, unlimited video hours, up to 10,000 hours per index — fits mid-size media and adtech teams whose monthly spend tracks actual footage. Enterprise committed-use fits large archives, public sector, and air-gapped deployments where published per-query rates don't hold.
In short
Twelve Labs — Video intelligence API that indexes, searches, and analyzes footage at scale with the Marengo and Pegasus models. Best for Media companies and archives with thousands of hours of footage, Adtech and contextual targeting platforms, Government and public sector evidence teams. Free to start; paid plans from $2.5.
What people actually say about Twelve Labs — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
16 mentions across 3 sources (Hacker News, YouTube, Lemmy) · researched Sep 1, 2026.
Average across the 3 sources that answered — each source counts once, not each post.
- +High-speed indexing: processes an hour of video in a minute.
- +Multimodal search across video, audio, and text (Marengo).
- +Natural language search across entire video libraries.
- +Supports 10,000+ hours per day on Developer tier.
- +Automatic scene segmentation with claimed Video-MME #1.
- −Reported as 'doesn't work' by some YouTube users.
- −Lack of third-party demos; only creator videos, raising suspicion.
- −Usage-based pricing can get expensive for smaller projects.
- −Not suitable for casual users; enterprise-first focus.
- −Unclear whether audio is fully resolved in search (question on Marengo).
- • Usage-based pricing can accumulate quickly, especially for small projects.
- • Possible additional costs for high-resolution video processing or custom integrations.
Viability Score
How well maintained and how widely used is Twelve Labs? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Natural language video search across full libraries
- Multimodal indexing of visual, audio, and speech signals
- Any-to-any retrieval (video, image, audio, text queries)
- Marengo 3.5 search and retrieval model
- Pegasus 1.5 video-to-text generation model
- Jockey video intelligence AI agent
- Automatic scene and pacing segmentation
- Compliance and brand safety detection with explainable AI
- Highlight creation and export into editing workflows
- Video insights generation at corpus scale
- Embed API for multimodal embeddings
- Search API availability through Jockey on Marengo 3.5
- Multimodal prompting with video, audio, image, and text inputs
- ~60x real-time processing, 10k+ hours per day
- SDKs for multiple languages
About Twelve Labs
Twelve Labs is a video intelligence platform that ingests, indexes, searches, segments, and generates text from video through an API. It runs two foundation models: Marengo 3.5 for multimodal search and any-to-any retrieval, and Pegasus 1.5 for video-to-text generation. It processes video at roughly 60x real-time speed — the vendor claims about an hour of video indexed per minute and 10,000+ hours per day. You can search an entire library with natural language (actions, scenes, dialogue, even human emotion) without manual tagging, auto-segment long-form footage at natural breaks, flag policy and brand-safety issues with explainable output, build highlight reels from descriptions, and generate structured metadata from unstructured video. Jockey, the vendor's video intelligence AI agent, is the intended front door to this — as of the current pricing page, the Search API is available through Jockey on Marengo 3.5. It is built for media and entertainment archives, sports, advertising and marketing, public sector evidence management, and security compliance work. It offers SDKs, MCP integrations, and air-gapped deployment, and is SOC 2 Type II certified. Free tier covers up to 10 hours of indexing; Developer is pay-as-you-go; Enterprise is committed-use.
Behind the Verdict
Twelve Labs sells infrastructure, not a point-and-click tool. The pitch is that a single pipeline ingests multimodal data, indexes it once, and then answers queries across every modality — you can search with text and retrieve video moments, or search with an image, with no tagging step. Three things make this more than a wrapper: the two proprietary foundation models (Marengo 3.5 for retrieval, Pegasus 1.5 for generation), the retrieval accuracy the vendor reports on Video-MME, and the operational scale (~60x real-time, 10k+ hours/day). That combination is hard to replicate with a prompt over an off-the-shelf model. The workflows on the homepage are concrete and worth understanding before you buy: Search & Discover (natural language across hours or years of footage), Segment Content (automatic scene and pacing breaks grounded in what happened on screen, not a transcript), Ensure Compliance (policy and brand-safety flags with explainable AI so a human reviewer can audit the call), Create Highlights (describe the cut you want — every scored goal this season — and it finds and assembles the material into an editing workflow), and Generate Insights (patterns across a corpus to inform creative decisions). Where it fits: media and entertainment archives that need timestamped clips from years of footage; sports teams pulling highlight packages; adtech and marketing teams doing contextual targeting against actual scene content instead of metadata; public sector and defense teams doing evidence management and anomaly detection, including air-gapped deployments. Where it doesn't: hobby projects, teams that want a no-code UI, or anyone with a small library where pay-as-you-go indexing at $2.50/hour per video hour is more expensive than the value retrieved. The vendor's own reported figures — 10x faster compliance review, 4-hour single video in one API call — are the kind of numbers worth piloting against your own footage rather than taking on faith.
Researching Twelve Labs? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Twelve Labs actually fits — and what changes day-one when you adopt it.
You have 3,000 hours of documentary footage spanning a decade. You index it once, then search for a described moment ('the argument on the boat at sunset') and pull timestamped clips straight into your edit.
Outcome: What used to take a research team three days takes seconds per query, per the vendor's archive workflow.
You run every piece of user-uploaded video through compliance detection and get explainable flags on policy risks and brand-safety issues instead of a black-box score.
Outcome: Reviewers scan flagged segments rather than full runtimes — the vendor claims 10x faster compliance scanning.
You build a contextual targeting layer: scenes are understood from visual and speech content, so ads only land in brand-safe segments without manual tagging or metadata rules.
Outcome: Serving decisions are driven by what actually happened on screen rather than tags that may be wrong.
Use Cases
- Search thousands of hours of surveillance footage for a specific event using a plain-language query.
- Generate highlight reels from sports games or event recordings based on a coach's or editor's description.
- Detect policy violations or brand safety issues in user-generated content with explainable alerts.
- Index a large video archive so production teams can find timestamped clips in seconds.
- Build a contextual advertising engine that places ads based on actual scene content rather than metadata.
- Produce structured metadata from unstructured video for compliance, legal, or archival use.
- Run video intelligence in air-gapped environments for government evidence management.
- Review long-form content for compliance about 10x faster than manual scanning.
Models Under the Hood
as of 2026-09-22
Limitations
- The Free tier caps at 10 hours of indexing, with 90-day index access, 10 hours per index, and 100 videos per index.
- The Developer tier is pay-as-you-go: video indexing is $2.50 per hour of video, Analyze API input video is $1.75 per hour, Search API is $4 per 1,000 queries, and output text is $7.50 per 1M tokens.
- On the Developer tier, duration per index is 10,000 hours and volume per index is 100,000 videos.
- Enterprise is committed-use with custom pricing.
- Rate limits apply on Enterprise and scale with monthly spend.
as of 2026-10-03
Verification history
We have re-verified Twelve Labs 8 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 8 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Twelve Labs tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0/mo
Ideal for
Developers and evaluators piloting video search on a small corpus — a handful of test videos or a single short production.
What this tier adds
Free entry point: up to 10 hours of indexing, free Search API and Analyze API usage, but index access expires after 90 days and each index is capped at 10 hours and 100 videos.
Developer
Pay as you go ($2.50/hour video indexing; $1.75/hour Analyze
Ideal for
Mid-size media, adtech, or product teams running a live video feature where monthly spend tracks actual footage processed.
What this tier adds
Adds unlimited video hours and index access, 10,000 hours per index, 100,000 videos per index, and 25 concurrent indexing tasks — in exchange for $2.50/hour video indexing, $4 per 1,000 Search API queries, and $1.75/hour Analyze input video.
Enterprise
Custom (committed use contracts)
Ideal for
Large archives, public sector, and regulated teams that need committed capacity, custom rate limits, and air-gapped deployment.
What this tier adds
Moves all Search, Analyze, Embed, and indexing rates to custom committed-use pricing with unlimited video hours and custom index durations and volumes.
Where the pricing makes sense
The company stage and team size where Twelve Labs's pricing actually pencils out — and where peers do it cheaper.
The Free tier (10 hours of indexing, 90-day index access) fits evaluation and pilot work. The Developer pay-as-you-go tier — $2.50 per hour of video indexed, $1.75/hour Analyze input video, $4 per 1,000 queries, unlimited video hours, up to 10,000 hours per index — fits mid-size media and adtech teams whose monthly spend tracks actual footage. Enterprise committed-use fits large archives, public sector, and air-gapped deployments where published per-query rates don't hold.
Setup time & first value
How long it actually takes to get something useful out of Twelve Labs — broken out by persona, not the marketing-page minute.
Free-tier evaluation is fast — sign up, index up to 10 hours, and run searches the same day. A Developer-tier production integration via the SDKs takes roughly a few days for a single pipeline; the real clock is indexing throughput, which the vendor puts at about an hour of video per minute of processing, so a 500-hour archive takes hours of wall-clock time rather than weeks.
Switching to or from Twelve Labs
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From manual tagging and MAM search: index your existing archive and replace tag-based lookup with natural-language queries and scene segmentation.
- →From Google Video Intelligence API: move from label detection to any-to-any retrieval, and re-index your footage into Twelve Labs indices.
- →From Google Gemini video understanding: the vendor publishes Pegasus 1.5 at +13.1% over Gemini 3.1 Pro on multimodal prompting, so benchmark the same clips before switching.
- →From an in-house CLIP-style retrieval stack: replace the embedding-and-search pipeline with Marengo 3.5 through the Embed and Search APIs.
- →From AWS Rekognition: keep your video in AWS, index through the API, and add speech-and-scene reasoning that labeling alone doesn't cover.
- ↗To Google Video Intelligence API: for small libraries where label detection is enough, a per-video cloud pricing model may undercut $2.50/hour indexing plus $1.75/hour Analyze.
- ↗To Gemini video understanding: if your workload is mostly prompting over a single video rather than retrieval across a corpus, a general-purpose model removes the separate indexing cost.
- ↗To an in-house open-source stack: teams with ML engineers can trade Twelve Labs' accuracy and scale claims for avoided per-hour indexing fees.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Twelve Labs”, and we withheld 6: 6 could not be judged, because “Twelve Labs” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Twelve Labs.
Official links
Tools that pair well with Twelve Labs
Common stack mates teams adopt alongside Twelve Labs, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Twelve Labs vs Spider Cloud
Choose Spider Cloud if your AI pipeline needs structured data from the web (especially for RAG) at scale with a cost of $0.03 per 1k pages and features like Browser AI commands, 1,000+ scraper examples, and data connectors. Choose Twelve Labs only if your primary data is video — it excels at search and analysis over thousands of hours of footage. For most AI agent use cases, Spider Cloud is the clear winner due to its lower cost, broader output formats, and more developer-friendly integrations.
Twelve Labs vs Voyage Ai
If your core need is high-accuracy text retrieval for enterprise RAG—especially with domain-specific finance, legal, or code data—Voyage AI's specialized embeddings and low-dimensional vectors offer a compelling, cost-efficient solution for large document collections. For organizations processing thousands of hours of video, Twelve Labs' multimodal Marengo/Pegasus models provide unrivaled any-to-any retrieval and analysis at 60x real-time speed. Choose based on your primary modality: text vs. video.
Twelve Labs vs Temporal Ai
Choose Temporal AI if your priority is building fault-tolerant, long-running workflows and AI agents that survive crashes; it's the leader in durable execution with flexible SDKs and strong enterprise adoption. Choose Twelve Labs if your core need is video intelligence at scale—searching, analyzing, and generating insights from massive video libraries using state-of-the-art multimodal models. They solve fundamentally different problems, so your decision should hinge on whether you need orchestration reliability or video understanding.
Alternatives to Twelve Labs
View all
Reka
Reka builds omni models for real-time video reasoning that run on-device, not just in the cloud.
Reka AI
Reka AI builds omni models, edge inference, and video infrastructure for the physical AI era.
PixelMotion
Turn product photos into professional marketing videos with 57 AI models, from $29/mo.
Frequently Asked Questions
Best-of guides
Used Twelve Labs? Help shape our editorial sentiment research.