Memories.ai - ChatGPT for your video library, with unlimited video context

I’m Shawn, founder of Memories.ai, former researcher at Meta, and PhD in Computer Science from the University of Cambridge.

🚀 Today we’re launching Memories.ai – the world’s first Large Visual Memory Model (LVMM). Our goal? Give AI human-like visual memory.

🧠 Why Visual Memory?

AI today is mostly chat-based. But humans aren’t just chat-based – we think, remember, and interact with the world visually. Visual memory is core to how we understand people, places, and context. That’s what we’re bringing to AI.

With Memories.ai, AI can see, understand, and remember your video content – just like a person would. Once your videos are uploaded and indexed, you can interact with them naturally, ask questions, and retrieve any moment instantly — no need to reprocess anything.

🎥 What You Can Do With Memories.ai

Chat with your video library: Ask questions and get answers from your video content, instantly.

Create videos via chat: Our Video Creator Agent (Beta) lets you generate video content simply by chatting.

• This is like Lovable for vidoe editing

• No more dragging and clipping

• Just one prompt, you get your edited video directly

Discover trends & influencers: Our Video Marketer Agent (Beta) indexes over 1 million TikToks so you can:

• Explore viral trends,

• Identify relevant creators — all in seconds.

• No more scrolling through endless feeds. Just chat and get what you need.

• Maybe perplexity for TikTok?

Thank you all for your support!

Wait, so I can actually *ask* my videos questions now? That’s wild—no more scrubbing through hours of footage. Super smart idea, fr. Props to the team!

Congrats! Is it possible to list your AI a few YT videos (let's say lectures and podcasts) and then ask it to organise some info from them into a single output?

Currently trying it (for finding a movie that I watched as a kid, but cannot find the right answer because I am bad at describing the scene). 😭

Anyway, the tool came up with the idea of creating the movie :D

I love video production and cinematography, so the tool sparked my attention immediately.

Can't wait to sign up; it's revolutionary for the workflow of video content creators.Can't wait to see how AI will evolve with visual memory. @Memories.ai

Congrats.. very cool porduct .. all the best

Congrats! Amazing Product！Top-Tier Team！

Congrats @imshawnshen on the launch. Will check it out.

wat

Congrats on launching Memories.ai! The ability for AI to remember and search videos like text is a game-changer!

You creating these for to remove the chat bot system between human and ai , but before creating the video we have to chat with AI right ??

This launch is fascinating! Memories.ai's ability to interact with video content just like a human is groundbreaking. I’m curious, how does the AI handle different languages in video content? Does it require subtitles for accurate understanding? 🤔

Ventsbusiness has built a strong name in online publishing by making it easy for new businesses and creators to share their stories.https://ventsbusiness.co.uk/ With flexible options like news releases and feature articles, the platform helps startups and PR teams get noticed without the hurdles of traditional media. This open and user-friendly approach supports fresh voices and gives readers a mix of real stories and smart branding.

What a cool product it is!

Congratulations on the launch of Memories.ai! I gave it a quick try by using it to analyze the creative angles in influencers' videos promoting products, and it clearly showed some promise.

For anyone attempting to figure out what truly works in that space, it helped bring to light a few patterns in their pitch structure that I hadn't previously been aware of.

Fantastic work so far! I'm eager to learn more and watch how this develops.

Ah! Saw you guys on X. And launched here too today.. amazing. Yet to check the product but sure it caught attention! Good luck.. @imshawnshen

Unlimited context is impressive! But:

1. Where is video processed? (On-device vs cloud)

2. How are temporal relationships stored? (e.g., 'Before X happened, Y occurred')

3. Can it distinguish identical objects across time? (e.g., 'Show all instances of this coffee mug')

Congrats on the launch!

This is pretty cool. It could be great for curating educational video content or helping content creators and YouTubers surface specific scenes from their older videos.

I wouldn’t be surprised if one of the big foundational model players tries to acquire this. A memory layer for video understanding is kind of a layer they want to embed into their existing models.