Understand long videos by decoupling perception and reasoning through hierarchical graph memory and an agentic retrieval mechanism.
SaferSkills independently audited MemDreamer (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
MemDreamer helps you analyze long video files. The software identifies key objects, actions, and events across hours of footage. It creates a compact digital memory of the video data. You can ask questions about the video content and receive accurate answers. This tool works by separating how the software sees images from how it understands logical links.
Before you install MemDreamer, ensure your computer meets these minimum standards:
Follow these steps to set up the software on your Windows computer.
.exe to your computer.Once you finish the installation, double-click the MemDreamer icon on your desktop to open the application.
Computers sometimes face minor errors. Review these tips if you encounter issues during operation.
MemDreamer processes your video files locally on your computer. Your files remain on your hard drive. The software sends text queries to the cloud to perform the reasoning tasks. It does not upload your actual video files to the internet. We prioritize the security of your media at all times.
MemDreamer uses two main methods to interpret your videos.
First, it uses a perception step. It scans every frame of your video to recognize objects and movements. It labels these items based on common patterns.
Second, it uses a reasoning step. It places these labels into a hierarchical graph memory. Think of this as a digital outline of the video. The software looks through this outline to connect events that happen far apart in time. This is how the system remembers things that occur early in a two-hour video when you ask a question at the very end.
Can I run this on a laptop? Yes, if your laptop meets the hardware requirements listed in the System Requirements section. Gaming laptops usually provide the necessary power.
How many videos can I analyze at one time? The software is designed to process one video at a time to ensure optimal performance and accuracy.
Is there a limit to the length of the video? There is no hard limit, but very long videos require more time for the initial analysis step.
Does the software work offline? The initial analysis of the video happens offline. You need an active internet connection to submit your questions for processing.
Will this software damage my files? No. The software only reads your video files. It does not perform any write or delete operations on your original source files.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.