Resources // Voice & Multimodal AI

Twelve Labs

Offer 600 free minutes of video understanding
Suits for Non-VC-backedResearchersCreators
Updated Jun 2026
Apply ↗

The offer

  • Free plan: 600 minutes of video (indexing + analyze + segment combined), up to 10 hours per index, 100 videos per index, 5 concurrent indexing tasks, no credit card
  • Free Search API, Embed API, and Pegasus analyze usage within that allotment
  • Developer (pay-as-you-go) above the free tier: Marengo indexing $0.042/min, Search $4 / 1,000 queries; Pegasus analyze $0.0292/min
  • Enterprise: custom contracts for scale, domain fine-tuning, and higher limits

Who qualifies

  • Open to anyone. First login auto-assigns the Free plan, no application
  • Self-serve upgrade to pay-as-you-go, no startup-specific discount (everyone is on the same public tiers)
  • Free indexes expire 90 days after creation, long-term storage of indexed video needs a paid plan
  • The 600 minutes are a finite shared budget across indexing and analysis, not 600 min/month forever

Community Insights

Twelve Labs is positioned as a leading video-understanding foundation-model provider (search, embeddings, and analysis over video), and external coverage frames it that way through funding news and cloud case studies. Independent hands-on discussion on developer forums is thin, so there is little verifiable community critique to report. The read is that the 600 free minutes make it cheap to evaluate whether multimodal video search fits your problem before you commit.

Best Practices (from community tips)

  • Index a representative slice of your real footage first to see how fast 600 minutes depletes for your resolution and length mix.
  • Separate the two models: Marengo (search / embed) and Pegasus (analyze / generate) are priced differently. Pick per task.
  • Export or migrate anything you need long-term before the 90-day free-index expiry.
  • Use the pricing calculator to project costs before scaling beyond the free tier.
  • For heavy, PII, or domain-specific video, raise Enterprise early. Fine-tuning and retention limits are negotiated there.