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

# v0.2.0 - 2026-06-25

NeMo Platform v0.2.0 expands the local-first OSS distribution with plugin
workflows for fine-tuning, synthetic data generation, and model auditing. It
also improves Docker-backed local workflows, SDK resources, and NAT-based agent
container tooling.

## Highlights

* **Fine-tuning.** Customizer is available through the plugin architecture,
  with Docker-backed Unsloth jobs, NeMo Automodel configuration support, vLLM
  adapter support, and updated tutorials.
* **Local model serving.** vLLM deployment in Docker supports local
  model-serving workflows for OSS development and testing.
* **Synthetic data generation.** Safe Synthesizer is available as a local
  plugin for privacy-preserving synthetic data generation workflows.
* **Model auditing.** Auditor adds local garak-backed audit configuration,
  target management, SDK resources, and documentation.
* **Agent containers.** `nemo agents` includes commands for rendering,
  building, and publishing agent containers.
* **Plugin platform.** Platform capabilities can ship as plugins while still
  using platform APIs, SDK resources, local services, and workspace scoping.

## What's included

### Platform and Plugins

* Plugin packaging for platform capabilities distributed outside the core
  package.
* Docker support for source-install local platform workflows.
* SDK and API support for plugin-provided resources.
* Source-install auth and OIDC bootstrap guidance for users who enable RBAC
  locally.

### Fine-tune Models

* Customizer migration to the plugin architecture.
* Docker-backed Unsloth SFT jobs.
* NeMo Automodel and Unsloth backend support for fine-tuning workflows.
* vLLM adapter support for serving trained adapters.
* Updated Customizer documentation and examples.

### Models and Inference

* vLLM deployment in Docker for local model-serving workflows.
* Local model-deployment documentation for the OSS quickstart and SDK usage.
* Continued Inference Gateway support for provider registration, virtual
  models, and OpenAI-compatible routing.

### Agents

* `nemo agents` container subcommands for rendering, building, and publishing
  agent containers.
* Continued NAT-based agent workflow support from the 0.1 line.

### Safe Synthesizer

* Safe Synthesizer packaged as a platform plugin.
* Local execution support for Safe Synthesizer workflows.
* Documentation and tutorials for data synthesis, PII replacement,
  evaluation, jobs, host-local development, and differential privacy.

### Auditor

* Auditor plugin APIs for audit configs and targets.
* Local garak-backed auditing workflows.
* SDK resources and documentation for configuring audits and reviewing
  results.

## Install

For a fresh local checkout:

```bash
git clone https://github.com/NVIDIA-NeMo/nemo-platform.git
cd nemo-platform
make bootstrap
source .venv/bin/activate
nemo setup
```

See [Setup](/documentation/get-started) for prerequisites and provider
configuration.

## Upgrade from v0.1.x

From an existing local checkout:

```bash
git fetch --tags
git checkout v0.2.0
make bootstrap
source .venv/bin/activate
nemo setup
```

If the `v0.2.0` tag is not available from GitHub yet, use the fresh checkout
flow above and return to the release tag after it is published.

After setup, restart local services before using CLI, SDK, Studio, or plugin
workflows against the upgraded checkout.

## Compatibility

* Python 3.11-3.14
* macOS and Linux for the OSS local install path
* Docker for Docker-backed platform, job, and vLLM model-serving workflows
* NVIDIA GPU access for local training, model serving, and GPU-backed
  synthetic data workflows
* Node 22.18.0+ for Studio assets
* Platform API `0.2.0` and `nemo-platform` Python SDK `0.2.0`

## Current constraints

* **Local-first OSS scope.** v0.2.0 improves local Docker-backed workflows. It
  is not a managed cluster or scale-out enterprise release.
* **Docker model serving.** v0.2.0 includes vLLM deployment in Docker. Docker
  deployment for NIM is not included in this release.
* **Customizer Docker support.** Docker-backed Customizer jobs target GPU
  Linux environments. ARM64 images and NeMo Automodel Docker execution are not
  part of this release.
* **Auth and RBAC.** Source-install auth and OIDC setups require the bootstrap
  IAM seed step described in [Setup](/documentation/get-started). Validate
  role bindings for your deployment mode before relying on them for multi-user
  access control.
* **Skill Evaluation.** Skill Evaluation workflows are not included in
  v0.2.0.

## Links

* Repository: [https://github.com/NVIDIA-NeMo/nemo-platform](https://github.com/NVIDIA-NeMo/nemo-platform)
* Issues: [https://github.com/NVIDIA-NeMo/nemo-platform/issues](https://github.com/NVIDIA-NeMo/nemo-platform/issues)
* NeMo Agent Toolkit: [https://docs.nvidia.com/nemo/agent-toolkit/latest/](https://docs.nvidia.com/nemo/agent-toolkit/latest/)