InfiniMind

Comprehensive brand identity guidelines for InfiniMind.

DeepFrame analyzes video footage to generate evidence-based insights for teams.

DeepFrame Safety enables natural-language video search and review for public-sector organizations within secure environments.

The platform analyzes TV broadcasts using AI to extract business insights.

Seed round led by UTEC to accelerate development of DeepFrame, a Large Video Model for enterprise-scale video understanding

InfiniMind won first place and $100,000 in OpenAI API credits at Series T - Post AGI from Kyoto, the OpenAI x Open Network Lab pitch contest held during IVS 2026.

An open benchmark measuring narrative and cultural reasoning in long-form Japanese video — and why current multimodal models still struggle.

Latest news and updates from InfiniMind.

Join InfiniMind and help build the infrastructure that turns video into intelligence.

InfiniMind hires sales engineers to translate customer needs into technical solutions and prototypes during meetings.

Engineers develop and deploy features across InfiniMind’s platform, from search interfaces to processing pipelines.

A detailed comparison of TVPulse's Starter, Standard, and Enterprise plans

InfiniMind converts video footage into actionable evidence for users.

InfiniMind offers video AI solutions for businesses.

InfiniMind provides help articles, guides, and FAQs for its products.

The intelligence layer for the physical world. Transform petabytes of video into structured, searchable intelligence.

InfiniMind Inc. outlines its privacy practices under Japan’s Act on the Protection of Personal Information.

InfiniMind's API Terms of Service governing use of the DeepFrame API.

InfiniMind's Terms of Service for DeepFrame.

InfiniMind secures first place in **Series T – Post AGI**, an OpenAI-backed pitch contest hosted by Open Network Lab in Kyoto. The event, part of IVS 2026 KYOTO, awarded the startup **$100,000 in OpenAI API credits**. The competition highlights innovative ventures addressing post-AGI challenges.

Jul 2026InfiniMindNews article

Infinimind introduces the NARU Bench, a new evaluation framework assessing multimodal models' ability to interpret high-context video content. The benchmark tests models on complex tasks like tracking shifting motivations in long interviews or decoding nuanced conversational cues, such as Japanese backchannels. It highlights current gaps in models' capacity to handle layered, real-world video understanding beyond simple factual retrieval.

Jun 2026InfiniMindNews article

InfiniMind, founded by former Google executives Aza Kai and Hiraku Yanagita, raises **$5.8 million** in seed funding to develop AI infrastructure for processing **enterprise video and audio data**. The round includes investors like UTEC, CX2, and an a16z Scout, with plans to expand operations by moving its headquarters to the U.S. while keeping a presence in Japan.

Feb 2026AI InsiderNews article

Former Googlers Aza Kai and Hiraku Yanagita co-founded Tokyo-based InfiniMind to turn large video and audio archives into queryable business data. The startup has secured $5.8 million in seed funding led by UTEC to develop its DeepFrame platform, expand engineering, and reach more customers in Japan and the U.S. Its TV Pulse product has paying media and wholesale customers, while DeepFrame is scheduled for beta in March and launch in April 2026.

Feb 2026TechCrunchNews article