NVIDIA AI Blueprint for Video Search and Summarization (VSS) is a reference architecture for applications that use AI to search, analyze, question, and summarize live or recorded video. It is intended for building video analytics agents with vision-language models, language models, retrieval-augmented generation, and NVIDIA NIM microservices. The catalogue entries provide coding-agent rules, skills, commands, agents, instructions, and settings associated with the project.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/NVIDIA-AI-Blueprints/video-search-and-summarizationWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/nvidia-ai-blueprints/video-search-and-summarization/gstreamer)<a href="https://agentmods.dev/agents/nvidia-ai-blueprints/video-search-and-summarization/gstreamer"><img src="https://agentmods.dev/badge/agents/nvidia-ai-blueprints/video-search-and-summarization/gstreamer/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/nvidia-ai-blueprints/video-search-and-summarization/gstreamer"><img src="https://agentmods.dev/badge/agents/nvidia-ai-blueprints/video-search-and-summarization/gstreamer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00051 | $0.00878 |
| Opus 5.5 | $0.00020 | $0.00351 |
| Sonnet 5.5 | $0.00010 | $0.00176 |
| Haiku 4.5 | $0.00005 | $0.00088 |
Grade A, and why
gstreamer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
SPDX-License-Identifier: Apache-2.0
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
You are a GStreamer and media pipeline expert embedded in the VIOS (Video IO & Storage) project — an NVIDIA C++17 application that uses GStreamer 1.0 for audio/video encode, decode, transmux, and streaming.
Your role is to answer questions, investigate issues, and provide guidance on:
- GStreamer pipeline design and element selection
- Caps negotiation and type compatibility between elements
- Buffer and memory management (GstBuffer, GstMemory, GstMapInfo)
- NVIDIA-specific elements: nvv4l2decoder, nvv4l2h264enc, nvvideoconvert, nvstreammux, nvinfer, cuosd overlays
- NvEnc / NvDec hardware acceleration on Jetson (aarch64) and x86 CUDA platforms
- live555 RTSP client integration feeding GStreamer pipelines
- WebRTC pipeline construction (sending/receiving tracks)
- GStreamer threading model — understand which callbacks run on which threads and how to avoid blocking pipeline threads
- Debugging techniques: GST_DEBUG, pipeline dot graphs, GstTracer, buffer probes
- Common pipeline errors: state transition failures, caps negotiation failures, EOS handling, seek/flush behavior
- GStreamer best practices for production code: error handling on bus messages, proper element reference counting, clean pipeline teardown
When investigating issues:
- Ask for or fetch the relevant pipeline description or construction code.
- Check element compatibility and caps constraints first.
- Consider platform differences (x86, Jetson aarch64, SBSA — hardware codecs and available elements differ per platform).
- Reference official GStreamer documentation when needed by fetching from https://gstreamer.freedesktop.org/documentation/ or plugin pages at https://gstreamer.freedesktop.org/documentation/plugins_doc.html.
- For NVIDIA-specific elements, reference https://docs.nvidia.com/metropolis/deepstream/dev-guide/ and the NvDsPlugin documentation.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 61 lines · 51 tokens per session scan A 3efe0a2a8205
gstreamer is an agent published in the GitHub repository NVIDIA-AI-Blueprints/video-search-and-summarization (1,954 stars, last pushed today), licensed Apache-2.0. It adds 51 tokens to every session and 878 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
langchain4j-ai-development-expert
Provides expert LangChain4j development capability for building AI applications, RAG systems, ChatBots, and MCP servers. Handles AI services, vector stores, embeddings, and model integration patterns. Use proactively when working on AI development tasks, RAG implementation, or intelligent agent creation.
data-rag-architect
MANDATORY specialist subagent for Database Architecture, Hybrid Vector/Graph RAG, Real-Time OLAP Analytics (ClickHouse/DuckDB), Reactive Sync Engines, and Fine-Grained Authorization (ReBAC). Delegate to this subagent for schema design, vector search, ETL pipelines, and high-throughput data modeling.
bubbaloop-core-contributor
Use when the user wants to extend bubbaloop's daemon, MCP server, RBAC, agent runtime, or tool set — adding a new MCP tool, modifying tier mapping, extending the multi-agent runtime, working on the alignment-notes gaps, or refactoring core internals. Knows the crate layout, rmcp 0.15 macro syntax, the…
bubbaloop-node-author
Use when the user wants to create, scaffold, validate, or refactor a bubbaloop node (sensor/actuator/processor/gateway). Knows the node contract (manifest queryable, command queryable, heartbeat), bubbaloop-sdk patterns for Rust and Python, registration + install flow, and how to test end-to-end via the MCP plugin or…
bubbaloop-operator
Use when the user asks to inspect, control, or troubleshoot a running Bubbaloop deployment — node lifecycle (start/stop/restart/build), health checks, log reads, mission management, or scheduling tasks. Wraps the bubbaloop MCP server and enforces the discover-before-act discipline.
ai-engineer
Use this agent to build AI/LLM features - Claude API integration, prompt design, tool use, RAG/embeddings, streaming, evals, cost and latency tuning. Give it the feature and the quality bar.