What if video AI had memory?
James Le (@le_james94) from @twelve_labs will touch on something that’s under-discussed.
Why do today's video systems reset with every query, and how can memory layers and context graphs enable a persistent, reusable understanding of video data?
TwelveLabs (twelvelabs.io)
2,548 posts
🎥 Building multimodal foundation models for video understanding: playground.twelvelabs.io
👉 Join our community: discord.gg/mwHQKFv7En
- In the 122nd session of #MultimodalWeekly, we will feature 2 exciting demo apps built with TwelveLabs from our recent Geospatial Video Intelligence hackathon in St. Louis. ✅ Bhavya Narula will present VoltHawk - which processes drone inspection footage through TwelveLabs
- TwelveLabs (twelvelabs.io) repostedWe hosted the coolest event yesterday at @CVPR 2026 with @ElorianAI @twelve_labs @NEA @NVIDIAAI See y’all at ICML! @Rxdha1 @thomastjoshi @nicoleegong @PikakeWin @gmi_cloud
- TwelveLabs (twelvelabs.io) repostedIn a @theDeepView exclusive, @twelve_labs first application-layer product, Rodeo, was featured showcasing how it enables #creatives to find, edit, and assemble footage using natural language prompts. Read more about what Ryan Khurana, TwelveLab’s Agents and Applications Lead, had
- TwelveLabs (twelvelabs.io) repostedI just gave my home security system a $200/month worth of upgrade for a $10/month price, all in 2 hours of honest work. 💦 And I open sourced the project: RingGuard 👇 github.com/saizo-ai/ringg… Here's how it works: 🏠 Run @home_assistant on an always-on computer in your home,
- Hosting our @Adobe Premiere Pro webinar next week: Brice Penven, Solutions Architect at TwelveLabs. He built this integration from the ground up - and on June 11th, he’ll show the full workflow live, from bin to rough cut without leaving the editing environment. Register now:
- The archive is full. The problem is finding anything in it. Mimir + TwelveLabs built a solution: semantic video intelligence inside your MAM. Natural language search, time-coded results, no new tools. Live demo June 18th. shorturl.at/KU8jq #TwelveLabs #VideoAI
- TwelveLabs (twelvelabs.io) repostedReplying to @Md_Sadiq_MdFor the implementation, I used @qdrant_engine for dense, sparse, and video vectors, @twelve_labs Pegasus and Marengo for video understanding and multimodal embeddings, and @llama_index to build an epoch-aware RAG layer on top of Relay The result is a replayable and verifiable
- TwelveLabs (twelvelabs.io) repostedVideo is the most information-dense modality we have, and most retrieval pipelines treat it like text with pictures. James Le from @twelve_labs is coming to Vector Space Day to show what's actually possible when you build multimodal retrieval right, from semantic search across
- TwelveLabs (twelvelabs.io) repostedSee how @twelve_labs just unveiled new tools to help fight AI slop by making it easier and faster for creators to produce and edit real video. This is an exclusive from The Deep View's Nat Rubio-Licht (@natrubio__). 🔗
- In the 121st session of #MultimodalWeekly, we will feature 3 exciting demo apps built with TwelveLabs from our recent Geospatial Video Intelligence hackathon in St. Louis. ✅ Kevin Lai will present Blocksight - which processes street-level video footage to identify storefronts
- Tomorrow: live demo of a broadcast pipeline that turns a live feed into social-ready short-form clips - automatically, before the broadcast is over. Register now. shorturl.at/ESzzY @quickplayott #TwelveLabs #VideoAI
- Most video teams do not have an editing problem. They have a search problem. Meet Rodeo, our new AI-powered creative copilot that helps creators go from raw footage to a first cut in minutes. Stop scrubbing, start directing. Try it here 👉 tryrodeo.io #VideoAI
00:00





















