Memories.ai
Category: Computer Vision
Memories.ai is an AI research lab developing a platform with long-term visual memory, allowing machines to analyze, understand, and search through millions of hours of video content. Memories.ai was founded in 2024. The company is led by Shawn Shen. Based in San Francisco, USA. Team size: 11-50. Total funding raised: $16M. Latest round: Seed. Key investors include Susa Ventures, Samsung Next, Crane Venture Partners, Fusion Fund, Seedcamp, Creator Ventures.
- Founded
- 2024
- Headquarters
- San Francisco, USA
- Team size
- 11-50
- Total funding
- $16M
Value proposition
To unlock insights from vast amounts of video data by providing a long-term memory for AI, enabling it to remember and contextualize information over extended periods, unlike current models that 'forget' after a few hours.
Products and solutions
Large Visual Memory Model (LVMM / LVMM 2.0), Video Intelligence API, Visual Search API, Project LUCI (on-device visual memory / LUCI Desktop & Pin), Ring camera visual memory integration, Agentic video analysis and summarization
Unique value
The company is pioneering the world's first Large Visual Memory (LVM) model, which mimics human memory patterns to overcome the limitations of current AI in long-context video analysis.
Target customer
Enterprises with large-scale video data, such as those in the security and media industries.
Industries served
Security & Safety, Media & Entertainment, Robotics & Physical AI, Consumer devices / wearables, Enterprise
Technology advantage
Their proprietary tech stack includes a multi-layer process for video analysis: noise removal, compression to store only important data, an indexing layer for natural language searchability with segmentation and tags, and an aggregation layer for summarizing data and creating reports. They also have a horizontal approach, allowing them to work with various video models.
How they differentiate
Memories.ai differentiates by focusing on creating a long-term visual memory for AI, enabling it to analyze and recall information from thousands or even millions of hours of video content. Competitors like mem0 and Letta are primarily focused on providing a memory layer for AI models with limited video support.
Main competitors
mem0, Letta
Key partnerships
NVIDIA, Qualcomm, Microsoft, Ring, Samsung, Lenovo, Vivo, Comcast
Notable customers
Ring (Appstore integration for camera visual memory), Named enterprise end-customers largely undisclosed; public technology partners include NVIDIA, Qualcomm, Microsoft, Samsung, Lenovo, Comcast
Major milestones
Oversubscribed $8M seed (Jul 2025) led by Susa Ventures; total funding to $16M with seed extension (reported Mar 2026), Launched world's first Large Visual Memory Model (LVMM); LVMM 2.0 with Qualcomm for on-device (Nov 2025), NVIDIA collaboration for robots, wearables, PCs (Oct 2025; GTC 2026), Project LUCI at Microsoft Build 2026 (on-device Windows ML visual memory), Ring Appstore integration (Apr 2026); SOC 2 Type II certified
Growth metrics
As an early-stage startup, specific growth metrics are not publicly available.
Market positioning
Memories.ai is positioning itself as a pioneer in Large Visual Memory (LVM) models. It targets enterprise applications across security, media, and retail industries that require analysis of extensive video footage.
Geographic focus
The company is headquartered in San Francisco, USA, suggesting an initial focus on the North American market.
Patents and IP
Information on patents is not publicly available.
About Shawn Shen
Former Research Scientist at Meta, PhD in Computer Vision from the University of Cambridge
More Computer Vision companies
Official website: https://memories.ai