Memories.ai and Qualcomm Collaboration.
Memories.ai Introduces Large Visual Memory Model 2.0 to Bring AI Visual Memory On-Device, in Collaboration with Qualcomm
Shannon Davis
4 days ago
Memories.ai today introduced its large visual memory model 2.0 (LVMM) to give AI systems visual memory on-device for the first time. As part of the announcement, memories.ai is announcing a collaboration with Qualcomm Technologies, Inc. to run natively on Qualcomm processors starting in 2026. The move will give consumers and businesses the ability to search and analyze visual moments securely and rapidly on their device.
Video’s Problem: Impossible to effectively search and analyze
Large language models (LLMs) made searching and analyzing text extremely fast and intelligent. It allowed consumers and enterprises the ability to search through hundreds of documents across different formats to quickly find and analyze text-based information. Video is different. It’s dense, unstructured, and stuck on a timeline. Teams still scrub hours of footage, rely on file names, or hope transcripts capture what the eyes see. Most tools index audio or thumbnails but miss actions, objects, and context. Without a visual memory layer, AI can’t find or recall the exact moments or answer broader questions in a sea of frames.
“By combining Qualcomm’s expertise in edge computing, connectivity, and on-device AI with Memories.ai’s innovative Large Visual Memory Model (LVMM), we are transforming how machines perceive, learn, and remember,” stated Vinesh Sukumar, VP, Product Management and Head of Gen AI/ML, Qualcomm Technologies, Inc. “This collaboration will enable AI platforms that are not only responsive but also context-aware, able to retain visual information, recognize patterns over long periods, and perform reliably even at the edge of networks. Together, we are speeding up our shared goal to deliver smarter, more intuitive intelligence to real-world applications”.
Memories.ai Solution: Large visual memory models that run on device, powered by Qualcomm
Memories.ai’s LVMM 2.0 solves these problems by turning raw video into structured memory on the device. It encodes frames, compresses them, and builds an index that supports sub-second search. Users can ask complex questions in plain language or use an image cue, and jump to the exact moment. Running natively on Qualcomm processors lowers latency, cuts cloud costs, and keeps data local for enhanced security. The model fuses video, audio, and images so results carry context, and a unified memory format keeps experiences consistent across phones, cameras, and edge systems. Developers get SDKs and reference designs to add capture, indexing, and retrieval without rebuilding backends. As more footage is indexed, LVMM improves recall and precision over time.
Real-world applications of LVMM 2.0 include:
- AI Albums: Organize and surface personal video memories on device.
- Smart Glasses and Wearables: Significantly enhanced AI recall.
- Security Agents: Help cameras understand and respond in real time.
- Robotics: Giving robots in the real world better context and understanding
“Qualcomm has proven themselves as a leader in bringing AI to the edge,” said Shawn Shen, Co-founder and CEO of Memories.ai. “We’re thrilled to partner with them to bring the incredible power of LVMM’s to hundreds of millions of phones, computers, and wearables in the coming years.”
PS
I use much stuff found here for ASML.
Return gift occasionally. |
