> This page is for version 1.0.0 (Latest) (default).
> For other versions, use one of these documentation indexes:
> - 1.0.0 (Latest) (default): https://docs.nvidia.com/nim/metadata-api/1.0.0/llms.txt

> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nim/metadata-api/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nim/metadata-api/_mcp/server.

# Ecosystem Overview

> Understand when to use tag-level metadata, NGC repository information, GPU lookup, and NIM deployment documentation.

The NIM Metadata API helps developers and automated tooling select profiles
within the NVIDIA NIM ecosystem before downloading containers. For API
capabilities, refer to [Overview](/nim/metadata-api/about/overview); for the publication and
request flow, refer to [Architecture Overview](/nim/metadata-api/about/architecture-overview).

## Understanding the NVIDIA NIM Ecosystem

NVIDIA NIM microservices package models with inference backends and deployable
profiles. The NGC Catalog identifies available containers and tags, while
metadata and repository APIs expose information used to evaluate them.

The NIM Metadata API's role is to return tag-level profiles, hardware
requirements, release details, and capabilities when those values were
recorded at publish time. It complements container-level catalog information
and the deployment documentation for each microservice.

## Choosing an Information Source

Choose the information source that matches your question:

| Source                                        | Use When                                                                 | Information Provided                                                                                |
| --------------------------------------------- | ------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------- |
| NGC Catalog and registry search               | Discover a NIM microservice and identify its repository and tags.        | Container identity and catalog discovery, including GPU filters.                                    |
| NIM Metadata API                              | Compare profiles for a specific microservice and tag before deployment.  | Tag-level hardware requirements, precision, backend, parallelism, and capabilities, when populated. |
| NGC repository endpoints                      | Retrieve information about a microservice as a whole.                    | Repository-level fields such as `name`, `nim_type`, `supported_models`, and `sdk_example_url`.      |
| NGC GPU metadata endpoint                     | Display a GPU name for a PCI device identifier.                          | Recorded subsystem variants and their `displayName` values.                                         |
| Microservice documentation and support matrix | Deploy the selected profile or investigate incomplete metadata coverage. | Microservice-specific deployment instructions and supported hardware.                               |

The metadata API and repository endpoints answer different questions. Make
both calls when you need tag-level profiles and container-level details. For
endpoint forms and authentication considerations, refer to
[Related Endpoints and Registry Search](/nim/metadata-api/reference/related-endpoints).

## How the Information Sources Work Together

Use the sources in the following order when planning a deployment:

1. Identify the NIM microservice and container tag in the NGC Catalog.
2. Query the NIM Metadata API for that repository and tag.
3. Compare available profiles against hardware and workload requirements.
4. Retrieve repository information or GPU display names when needed.
5. Follow the microservice's deployment instructions with the selected `profile_id`.

Metadata access and container entitlement are separate. Reading metadata does
not grant access to pull the container. If metadata is unavailable, consult
the microservice's documentation rather than assuming that the container is
unavailable or unsupported.

## Next Steps

Continue with the resource that matches your goal:

#### [Make Your First Request](/nim/metadata-api/1.0.0/get-started/quickstart)

Query a microservice and read the metadata response.

#### [Select a Profile](/nim/metadata-api/1.0.0/guides/choose-a-profile)

Apply hardware constraints and record a profile identifier.

#### [Query Complementary Endpoints](/nim/metadata-api/1.0.0/reference/related-endpoints)

Retrieve container information and resolve GPU identifiers.

#### [Handle Missing Metadata](/nim/metadata-api/1.0.0/troubleshooting)

Diagnose status codes and unrecorded requirements.