InfiniMind, a Tokyo-founded startup led by ex-Google Japan executives Aza Kai and Hiraku Yanagita, launched TV Pulse in April 2025 and has introduced DeepFrame as its long-form video intelligence product, featuring AI that converts large video and audio archives into searchable business data.
The company built infrastructure to process petabytes of footage, designed to identify scenes, speakers and events across unlimited-length files and to integrate visual, audio and speech signals. InfiniMind raised $5.8 million in seed funding and relocated its HQ to the U.S. while keeping a Japan office; DeepFrame entered beta in March 2026 with a planned full launch in April 2026.
For enterprises, the platform turns previously unused “dark” video data into queryable insights for monitoring, PR measurement and security, offering a no-code, cost-conscious route to large-scale media intelligence.
Image Credit: InfiniMind
Why This Trend Is Growing
- Long-form Video Intelligence
- Platforms that index and query unlimited-length video archives enable extraction of temporal insights previously locked in hour-scale footage.
- Multimodal Signal Fusion
- Combining visual, audio, and speech signals into unified queries creates richer context-aware analytics across disparate media types.
- Dark Data Monetization
- Transforming unused enterprise video and audio repositories into searchable business data reveals latent value streams for analytics and benchmarking.
Industries Being Reshaped
- Media Monitoring
- Advanced indexing of long-form broadcasts and recordings supports granular measurement of content reach, sentiment, and event-level impact.
- Corporate Security and Compliance
- Searchable video archives with speaker and event detection increase visibility into incidents, policy adherence, and forensic review capabilities.
- Public Relations and Brand Measurement
- Queryable media intelligence across extensive footage offers deeper attribution of brand mentions and campaign effectiveness over time.
