#6060th of 80 in Agent platforms

video-search-and-summarization

NVIDIA reference stack for searching and summarizing video with VLMs

Stars
2k
License
Apache-2.0
Last commit
Nov 2026
Last release
Oct 2026
Language
Python
Stars, last 29 weeks+1,376
21 Apr 2026: 598 stars28 Apr 2026: 607 stars5 May 2026: 617 stars12 May 2026: 629 stars19 May 2026: 1,360 stars26 May 2026: 1,427 stars2 Jun 2026: 1,473 stars9 Jun 2026: 1,511 stars16 Jun 2026: 1,525 stars23 Jun 2026: 1,548 stars30 Jun 2026: 1,663 stars7 Jul 2026: 1,690 stars14 Jul 2026: 1,711 stars21 Jul 2026: 1,729 stars28 Jul 2026: 1,736 stars4 Aug 2026: 1,758 stars11 Aug 2026: 1,780 stars18 Aug 2026: 1,796 stars25 Aug 2026: 1,812 stars1 Sept 2026: 1,828 stars8 Sept 2026: 1,839 stars15 Sept 2026: 1,856 stars22 Sept 2026: 1,868 stars29 Sept 2026: 1,874 stars6 Oct 2026: 1,905 stars13 Oct 2026: 1,922 stars20 Oct 2026: 1,932 stars27 Oct 2026: 1,949 stars3 Nov 2026: 1,961 stars10 Nov 2026: 1,974 stars

598 on 21 Apr 20261,974 on 10 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Reference implementation of NVIDIA's Video Search and Summarization blueprint. It includes microservices for RTSP/ONVIF ingest, real-time VLM captioning and embeddings, behavior analytics, alert verification and long-video summarization, plus an MCP-based agent. It uses NVIDIA NIM models (Cosmos3 Nano Reasoner, Nemotron 3.5 Lightning 30B A3B) and deploys with Docker Compose, Helm or a Brev launchable.

Who it is for: Engineers building video analytics agents on NVIDIA GPUs

Strengths

  • Covers ingest, embeddings, analytics, alert verification and summarization in one repository
  • Agent exposes video tools over MCP; ships agentskills.io-compatible skills
  • Separate reference workflows for Q&A, alerts, search and long video summarization
  • Documents validated OS, driver and Docker versions, including Jetson and DGX targets

Weaknesses

  • Tied to NVIDIA GPUs, NIM microservices and NGC-hosted images
  • Hosting NIM locally requires an NVIDIA AI Enterprise developer license
  • Video embedding search workflow is marked alpha
  • Many services (Elasticsearch, Redis, Kafka) make setup heavy; README says it needs technical expertise

What it needs

  • GPU required
  • Compose
  • Needs NVIDIA NIM, NGC API key, Docker Engine 28.3.3 to below 29.5.0, Docker Compose v2.39.1+, NVIDIA Container Toolkit 1.17.8+, Elasticsearch, Redis, Kafka
  • Models: Cosmos3 Nano Reasoner, NVIDIA Nemotron 3.5 Lightning 30B A3B

Quick answers

Is video-search-and-summarization open source?

Yes. Its code is on GitHub under the Apache-2.0 license.

Can you run video-search-and-summarization with Docker?

Yes. Its repository has a Docker setup: a Dockerfile, a Compose file or a published image.

Does video-search-and-summarization need a GPU?

Yes, it needs a GPU.

Also in Agent platforms

See all 80
Also in Agent platforms
RankProjectScore
1AutoGPTRole-based AI agents coordinated by one lead agent, plus a visual workflow builder Live demo ↗ (opens in a new tab)188k stars+853 (+0.5%) in 30 dayscustom license95 out of 100
2daily_stock_analysisLLM-driven daily analysis of watchlist stocks across six markets, pushed to chat apps69.9k stars+3,683 (+5.6%) in 30 daysMIT95 out of 100
3LangflowVisual builder for AI agents and workflows, exposed as API or MCP server156.5k stars+1,251 (+0.8%) in 30 daysMIT94 out of 100
#4deer-flowAgent harness running sub-agents, memory and sandboxes with pluggable skills85.2k stars+1,752 (+2.1%) in 30 daysMIT93 out of 100
#5DifyVisual builder for LLM workflows, RAG pipelines and agents, with APIs Live demo ↗ (opens in a new tab)161k stars+3,111 (+2.0%) in 30 dayscustom license92 out of 100