Many AI tools today can summarize individual videos, but understanding and analyzing multiple videos or footage spanning hours remains a challenge. This limitation affects security companies monitoring thousands of hours of camera footage and marketing firms evaluating various video campaigns.
Memories.ai is developing an AI platform capable of processing up to 10 million hours of video. The startup provides a contextual layer for companies with large video libraries, offering searchable indexing, tagging, segmentation, and data aggregation to make sense of vast amounts of footage.
Co-founder Dr. Shawn Shen, a former Meta Reality Labs researcher, and Enmin (Ben) Zhou, a former Meta machine learning engineer, founded the company to create a system that mirrors how humans use visual memory to sift through extensive data. “All top AI companies like Google, OpenAI, and Meta focus on end-to-end models,” Shen explained. “But these models often struggle with video context beyond one or two hours. We want to build a solution that better understands video across many hours.”
Funding and Support from Leading Investors
Memories.ai recently raised $8 million in a seed round led by Susa Ventures, with participation from Samsung Next, Fusion Fund, Crane Ventures, Seedcamp, and Creator Ventures. Interest from investors exceeded initial expectations, reflecting confidence in the company’s potential.
Misha Gordon-Rowe from Susa Ventures noted the startup’s ability to unlock vast amounts of visual intelligence data and identified a gap in long-context video understanding. Samsung Next’s Sam Campbell highlighted the platform’s potential for consumer applications, especially given its capacity for on-device computing, which could alleviate privacy concerns around cloud storage for home security footage.
The platform removes noise and compresses video data to store only essential information. It then indexes the video content for natural-language, searchable queries and adds segmentation and tagging to organize the clips. Finally, it aggregates the indexed data to generate insightful reports.
Market Focus and Future Plans
Currently, Memories.ai serves marketing and security sectors. Marketing teams can analyze social media video trends and receive guidance on content creation, while security companies use the platform to detect potentially dangerous activities by analyzing surveillance footage patterns.
Clients upload video libraries for analysis, but the company plans to introduce shared drives for easier content synchronization. Future features aim to allow users to query their videos with simple commands like, “Tell me all about people I interviewed last week.” Shen also sees potential applications in smart glasses, humanoid robot training, and self-driving car navigation memory.
Memories.ai competes with startups like mem0 and Letta, which offer memory layers for AI but with limited video support. Larger companies such as TwelveLabs and Google also work on video understanding, but Shen believes Memories.ai’s horizontal approach allows it to integrate effectively with various video models.
Author’s Opinion
Memories.ai’s approach to tackling the complex challenge of long-duration video analysis is impressive and promises to open new possibilities for industries like security and marketing. However, with the increasing reliance on AI for surveillance and personal data, it is crucial that privacy and ethical considerations remain front and center. The balance between powerful AI capabilities and protecting individual rights will define the success and acceptance of such technologies.
Featured image credit: Pierre Lecourt via Flickr
For more stories like it, click the +Follow button at the top of this page to follow us.