NGC Catalog
CLASSIC
Welcome Guest
Collections
Morpheus

Morpheus

For contents of this collection and more information, please view on a desktop device.
Features
Description
NVIDIA Morpheus enables cybersecurity developers and ISVs to build incredibly performant pipelines for security workflows with minimal development effort.
Curator
NVIDIA
Modified
July 29, 2025
Containers
Sorry, your browser does not support inline SVG.
Helm Charts
Sorry, your browser does not support inline SVG.
Models
Sorry, your browser does not support inline SVG.
Resources
Sorry, your browser does not support inline SVG.

Overview

NVIDIA Morpheus is a GPU-accelerated cybersecurity AI framework that makes it easy to build and scale cybersecurity applications that harness adaptive pipelines supporting a wider range of model complexity than previously feasible. Morpheus makes it possible to analyze up to 100 percent of your data in real-time, for more accurate detection and faster remediation of threats as they occur. Morpheus also provides the ability to leverage AI to adjust to threats and compensate on the fly, at line rate. With Morpheus organizations can attack the issue of cybersecurity head on. Rather than continuously chasing the cybersecurity problem, Morpheus provides the ability to propel you ahead of a breach and address the cybersecurity issue. With the world in a "discover and respond" state, where companies are finding breaches much too late, in a way that is way behind the curve, NVIDIA’s Morpheus cybersecurity AI framework enables any organization to warp to the present and begin to defend itself in real time.

  • Massive Performance and Scale - Morpheus is GPU-accelerated enabling, for the first time, the ability to inspect all network traffic in real-time, flag anomalies, and provide insights on these anomalies so that threats can be addressed quickly. It enables AI inference and real-time monitoring of every server and packet across the entire network.

  • Rapid Development and Deployment - Morpheus integrates AI frameworks and tools that make it easier for developers to build cybersecurity solutions. Organizations that lack AI expertise can still leverage AI for cybersecurity because Morpheus leverages tools for every stage of the AI workflow, from data preparation to training, inference, and deploying at scale. 

  • Real-time Telemetry - The Morpheus native graph streaming engine can receive rich, real-time network telemetry from every NVIDIA BlueField DPU-accelerated server or NVIDIA AppShield in the data center without impacting performance. Integrating the framework into a third-party cybersecurity offering brings the world’s best AI computing to communication networks. 

  • AI Cybersecurity Capabilities – Deploy your own models using common deep learning frameworks. Or use a Morpheus pre-trained and tested model to get a jump-start in building applications to identify leaked sensitive information, detect malware or fraud, do network mapping, flag user behavior changes, or and identify errors via logs.

Morpheus Framework

What's New: Morpheus 25.06.00 (23 Jul 2025)

🚨 Breaking Changes

  • Remove NeMoLLM as a dependency, and remove as an option in LLM pipelines (#2272) @dagardner-nv
  • Adopt Python 3.12 (#2184) @dagardner-nv
  • Bugfix: Remove MORPHEUS_USE_CONDA option; Add warning if CONDA_PREFIX not found (#2171) @willkill07

🐛 Bug Fixes

  • Fix: typo for PROJECT_IS_TOP_LEVEL for building examples (#2291) @willkill07
  • Fix mlflow dependency in Conda packages (#2290) @dagardner-nv
  • Fix devcontainer builds (#2288) @dagardner-nv
  • Fix type checking of the input message in PassThruStage example (#2284) @dagardner-nv
  • Replace default model in RAG pipeline with Nvidia hosted model (#2283) @dagardner-nv
  • Allow empty DataFrame slices (#2282) @dagardner-nv
  • Fix Sid visualization example (#2278) @dagardner-nv
  • Fix: examples build on aarch64 (#2277) @willkill07
  • Remove NeMoLLM as a dependency, and remove as an option in LLM pipelines (#2272) @dagardner-nv
  • Fix incompatible pynvml versions in conda envs (#2264) @dagardner-nv
  • Fix DFP docker build issues (#2261) @dagardner-nv
  • Fix handling of labels in DLP example (#2257) @dagardner-nv
  • Add optional databricks packages to the all env, and rename the group (#2251) @dagardner-nv
  • Support running the DLP example within the release container (#2244) @dagardner-nv
  • Clamp marshmallow to <4 to avoid version incompatibilities (#2235) @dagardner-nv
  • Fix conda package dependency issues (#2229) @dagardner-nv
  • Work-arounds for dependency issues (#2225) @dagardner-nv
  • Fix script reference in DFP example's Dockerfile from 125 to 128 (#2215) @Misiac
  • Fix docker container for Python 3.12 (#2208) @dagardner-nv
  • Invoke CudfHelper::load() from the main Python thread (#2207) @dagardner-nv
  • Add test to verify that #2202 is fixed (#2206) @dagardner-nv
  • Pin cupy to version <13.4; fix documentation link (#2194) @willkill07
  • Bugfix: Remove MORPHEUS_USE_CONDA option; Add warning if CONDA_PREFIX not found (#2171) @willkill07
  • Several fixes & improvements of C++ MonitorStage (#2170) @yczhang-nv
  • Bugfix: update old references of 25.02 to 25.06 (#2151) @willkill07

📖 Documentation

  • Update nlp_si_detection example README.md to reflect that boolean values are used in the output (#2293) @dagardner-nv
  • Update known issues document (#2292) @dagardner-nv
  • Update DLP diagram (#2255) @hsin-c
  • Update examples content list README.md to include DLP pipeline in the list (#2250) @tzemicheal
  • Update DLP README and notebook to include new diagrams (#2246) @hsin-c
  • Document the need to build the DLP example with the MORPHEUS_BUILD_EXAMPLES=ON flag (#2243) @dagardner-nv
  • DLP diagram formatting update (#2240) @hsin-c
  • Fix incorrect and outdated comments in examples documentation (#2234) @dagardner-nv
  • Add analytics to rendered documentation (#2205) @dagardner-nv

🚀 New Features

  • DLP inference example notebook (#2227) @tzemicheal
  • Add Data Loss Prevention Example (#2226) @dagardner-nv
  • Switch to the Nvidia Sphinx Theme for documentation (#2203) @dagardner-nv
  • Adopt Python 3.12 (#2184) @dagardner-nv
  • Adopt Rapids 25.02 and CUDA 12.8 (#2183) @dagardner-nv
  • Kinetica Vector DB Service (#2098) @am-kinetica

🛠️ Improvements

  • Set a reasonable num_threads default value for ransomware_detection example (#2268) @dagardner-nv
  • Update Dependency (#2230) @Salonijain27
  • Remove cudf_helpers.pyx patch for 24.12 bug (#2201) @mroeschke
  • Forward merge branch-25.02 into branch-25.06 (#2197) @willkill07
  • CI: Add new L4 runner for test (#2145) @willkill07

License

Morpheus is distributed as open source software under the Apache Software License 2.0.

NVIDIA AI Enterprise

NVIDIA AI Enterprise provides global support for NVIDIA AI software, including Morpheus. For more information on NVIDIA AI Enterprise please consult this overview and the NVIDIA AI Enterprise End User License Agreement.