# NVIDIA FLARE

> NVIDIA FLARE is a domain-agnostic, open-source, extensible SDK for federated learning and other federated-computing applications, from local development through production deployment.

## Overview

- [Welcome](welcome.md): Understand federated computing, federated learning as one application, NVIDIA FLARE, data locality, and the development-to-production lifecycle.
- [What's New](whats_new.md): Review current release features and changes.
- [Industry Use Cases](industry_use_cases.md): Explore applications and supporting material across industries.

## Get Started

- [Quick Start](quickstart.md): Run and understand a complete two-client federation, then choose the next path.
- [Installation](installation.md): Install NVIDIA FLARE and set up development or runtime environments.
- [Migration Guide](migration_guide.md): Upgrade between NVIDIA FLARE versions and understand behavior changes.

## Adapt And Build

- [Agent Skills](user_guide/agent_skills/index.md): Use maintained coding-agent workflows for supported training projects, federated statistics, optimization, and diagnosis.
- [Choose an API Path](user_guide/data_scientist_guide/api_selection.md): Select the Client API, Recipe, Collaboration API (Technical Preview), or lower-level components for a manual or custom integration.
- [Client API](user_guide/data_scientist_guide/client_api_usage.md): Use the high-level API for federated training scripts.
- [Getting Started with Recipes](user_guide/data_scientist_guide/job_recipe.md): Follow a task-oriented tutorial to create and run a job with a concrete recipe.
- [Recipe API Reference](user_guide/data_scientist_guide/recipe_api.md): Look up method signatures, helper behavior, ordering requirements, and guarantees for those same recipes.
- [Available Recipes](user_guide/data_scientist_guide/available_recipes.md): Browse bundled job recipes and supported patterns.
- [Collaboration API (Technical Preview)](user_guide/data_scientist_guide/collab_api.md): Design custom server-controlled or peer-to-peer algorithms and workflows.
- [FLARE API](user_guide/data_scientist_guide/flare_api.md): Use runtime APIs to submit, monitor, and manage jobs.
- [CLI Tools](user_guide/nvflare_cli/nvflare_cli.md): Use NVFlare command-line tools for local and deployed systems.

## Capabilities

- [Federated Statistics](examples/federated_statistics_overview.md): Compute aggregate statistics while source records remain at participating sites.
- [Federated LLM Fine-Tuning](programming_guide/llm_fine_tuning.md): Fine-tune and evaluate large models across sites.
- [Hugging Face Client API](user_guide/data_scientist_guide/hf_client_api.md): Adapt Transformers Trainer or TRL SFTTrainer scripts for federated training.
- [Federated XGBoost](user_guide/data_scientist_guide/federated_xgboost/federated_xgboost.md): Train XGBoost models in federated settings.
- [Mobile Training](user_guide/edge_development/mobile_training.md): Build mobile and hierarchical federated-learning applications.

## Examples, Tutorials, And Research

- [Example Applications](example_applications_algorithms.md): Find runnable applications, algorithms, and reference implementations.
- [Tutorials](tutorials.md): Learn from tutorials and notebook-based walkthroughs.
- [Self-Paced Training](self-paced-training/index.md): Follow the versioned multi-part training curriculum.
- [Research Papers](user_guide/researcher_guide/index.md): Browse published research and reference implementations.

## Deployment And Operations

- [Deployment Overview](user_guide/admin_guide/deployment/overview.md): Plan and run production NVIDIA FLARE deployments.
- [Provisioning System](programming_guide/provisioning_system.md): Provision participants, startup kits, and deployment assets.
- [Distributed Provisioning](user_guide/nvflare_cli/distributed_provisioning.md): Use distributed provisioning workflows.
- [Preflight Check](user_guide/nvflare_cli/preflight_check.md): Validate system readiness before launch.
- [System Configuration](user_guide/admin_guide/configurations/system_configuration.md): Configure runtime system behavior.
- [Monitoring](user_guide/admin_guide/monitoring.md): Monitor running federated learning systems.

## Security And Compliance

- [Security Overview](system_architecture/security_overview.md): Understand authentication, authorization, privacy, and auditing.
- [Identity And Access Control](user_guide/admin_guide/security/identity_security.md): Configure identity and access controls.
- [Communication Security](user_guide/admin_guide/security/communication_security.md): Secure network communication.
- [Data Privacy Protection](user_guide/admin_guide/security/data_privacy_protection.md): Protect data with filters and privacy mechanisms.
- [Differential Privacy](user_guide/admin_guide/security/differential_privacy.md): Apply differential privacy in federated learning.
- [Confidential Computing](user_guide/confidential_computing/index.md): Deploy with hardware-backed trusted execution environments.

## Developer Reference

- [Developer Guide](developer_guide.md): Understand architecture, controllers, filters, and extension points.
- [API Reference](apidocs/modules.md): Browse generated Python API documentation.
- [Glossary](glossary.md): Look up NVIDIA FLARE terms and concepts.
- [Contributing](contributing.md): Build docs, run tests, and contribute changes.