# NVIDIA/NeMo-Agent-Toolkit

The NVIDIA NeMo Agent toolkit is an open-source library for efficiently connecting and optimizing teams of AI agents.

Repository: https://github.com/NVIDIA/NeMo-Agent-Toolkit
Canonical: https://ross.abutalabs.com/products/nemo-agent-toolkit
Homepage: https://docs.nvidia.com/nemo/agent-toolkit/latest/
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-08-23T22:59:02+00:00

## Health v2 (maintenance only)
Score: 83/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 89, longevity 38
- inputs: {"age_days": 545, "days_push": 10, "days_rel": 78, "gap_med": 27, "n_releases_24m": 14}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2603, forks 747 (observed 2026-08-28T04:07:03.460317+00:00)

## What it is
NVIDIA NeMo Agent Toolkit is an open-source Python library for connecting, orchestrating, and optimizing teams of AI agents across any agentic framework such as LangChain, LlamaIndex, CrewAI, and Semantic Kernel. It provides YAML-configured workflows, profiling, evaluation, observability integrations, and serving via CLI commands or MCP/A2A servers.

## Use cases
- connect existing enterprise agents to tools and data sources across frameworks
- profile multi-agent workflows to find bottlenecks and token costs
- evaluate and optimize agent prompts and parameters automatically
- serve an agent workflow as an HTTP or MCP server
- build multi-agent systems that interoperate via the A2A protocol
- run agents against NVIDIA NIM, OpenAI, Azure OpenAI, or AWS Bedrock models

## When to choose
- you need a framework-agnostic layer to unify agents built with different frameworks
- you want enterprise-grade profiling, evaluation, and observability for agent workflows
- you are invested in the NVIDIA ecosystem (NIM, NeMo Retriever) or need A2A/MCP serving
- you want YAML-driven, composable agent workflows reusable across applications

## When to avoid
- you want a lightweight single-agent framework with minimal setup
- you need a framework with the largest community and third-party ecosystem
- your use case is simple prompting or chat without multi-agent orchestration
- you prefer not to depend on NVIDIA-maintained tooling

## Facets
- artifact type: library
- maturity: active
- function: agent-framework, llm-inference, rag, mcp, monitoring, workflow-automation
- domain: large-language-models, artificial-intelligence, developer-tools
- platform: python, cloud, self-hosted
- tags: multi-agent, observability, profiling, framework-agnostic, nvidia, yaml-config, a2a-protocol, evaluation, ai-agents, docker

## Member repositories
- NVIDIA/NeMo-Agent-Toolkit (main) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:03.460317+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:21:06.804298+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIA/NeMo-Agent-Toolkit (fetched 2026-08-28T04:07:03.460317+00:00, sha 553e92a3ba1e)
  - homepage: https://docs.nvidia.com/nemo/agent-toolkit/latest/ (fetched 2026-08-29T10:04:10.113506+00:00, sha 2a9d5ea31ca6)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/get-started/installation.html (fetched 2026-08-29T10:04:10.122831+00:00, sha c54e5bb89a39)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/build-workflows/about-building-workflows.html (fetched 2026-08-29T10:04:10.125213+00:00, sha 460e954b9bf6)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/run-workflows/about-running-workflows.html (fetched 2026-08-29T10:04:10.127189+00:00, sha 5077cb7fac6e)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/improve-workflows/about-improving-workflows.html (fetched 2026-08-29T10:04:10.129400+00:00, sha 75c7162ba13e)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/components/integrations/index.html (fetched 2026-08-29T10:04:10.131382+00:00, sha 6d4b80468db3)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/components/integrations/frameworks.html (fetched 2026-08-29T10:04:10.133212+00:00, sha 55413c162d5a)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/components/integrations/a2a.html (fetched 2026-08-29T10:04:10.135738+00:00, sha 5e5d3130dadc)
  - site_page: https://docs.nvidia.com/nemo/agent-toolkit/latest/components/integrations/integrating-aws-bedrock-models.html (fetched 2026-08-29T10:04:10.137706+00:00, sha f79633b6260a)
- Data as of 2026-08-30T08:39:29.467469+00:00.