Project G-Assist is an experimental on-device AI Assistant that helps RTX users control a broad range of PC settings, from optimizing game and system settings, charting frame rates and other key performance statistics, to controlling select peripheral lighting — all via basic voice or text commands.
Project G-Assist is built for community expansion. Whether you're a Python developer, C++ enthusiast, or just getting started — its Plugin architecture makes it easy to define new commands for G-Assist to execute. We can't wait to see what the community dreams up!
- Leverage a responsive Small Language Model (SLM) running locally on your own RTX GPU
- Extend and customize G-Assist with functionality that enhances your PC experience
- Interact with G-Assist from the NVIDIA Overlay without needing to tab out or switch programs
- Invoke AI-powered GPU and system controls in your applications using C++ and python bindings
- Integrate with agentic frameworks using tools like Langflow to embed G-Assist in bigger AI pipelines
- Python plugins for rapid development
- C++ plugins for performance-critical applications
- AI-driven features using the ChatGPT-powered Plugin Builder
- Custom system interactions for hardware and OS automation
- Game and application integrations that enhance PC performance or add new commands
If you're looking for inspiration, check out our sample plugins for controling peripheral & smart home lighting, invoking larger AI models like Gemini, managing Spotify tracks, or even checking streamers' online status on Twitch — and then let your own ideas take G-Assist to the next level!
Get started quickly using our Python bindings of the C++ APIs:
- Install the binding locally
cd api/bindings/python
pip install .- Chat with G-Assist
from rise import rise
# Initialize G-Assist connection
rise.register_rise_client()
# Send and receive messages
response = rise.send_rise_command("What is my GPU?")
print(f'Response: {response}')
"""
Response: Your GPU is an NVIDIA GeForce RTX 5090 with a Driver version of 572.83.
"""- Extend G-Assist
💡 Requirements:
- Python 3.x
- G-Assist core services installed
- pip package manager
See our Python Bindings Guide for detailed examples and advanced usage.
Try these commands:
- "Hey Twitch, is Ninja live?"
- "Check if shroud is streaming"
- "Is pokimane online right now?"
When a streamer is live:
ninja is LIVE!
Title: Friday Fortnite!
Game: Fortnite
Viewers: 45,231
Started At: 2024-03-14T12:34:56Z
When a streamer is offline:
ninja is OFFLINE
- Secure API credential management
- OAuth token handling
- Comprehensive logging system
- Windows pipe communication
- Real-time stream status checking
plugins/twitch/
├── manifest.json # Plugin configuration
├── config.json # Twitch API credentials
├── plugin.py # Main plugin code
└── requirements.txt # Dependencies (requests)
See our Twitch Plugin Example Code for a step-by-step guide to creating a Twitch integration plugin for G-Assist.
- Project G-Assist Plugins
- Why Plugins Matter
- What Can You Build?
- Quick Start
- G-Assist Module Architecture
- Extending G-Assist (Plugins)
- NVIDIA-Built G-Assist Plugins
- Community-Built Plugins
- Development Tools
- Need Help?
- License
- Contributing
flowchart TD
A[System Assist Module]
A -->|Runs Inference| B[Inference Engine]
A -->|Implements Built In Functions| C[Core Functions]
A -->|Launches| D[Plugin Launcher]
D --> E[Plugin 1]
D --> F[Plugin 2]
D --> G[Plugin n]
H[Community Code]
H -->|Develops & Contributes| D
Transform your ideas into powerful G-Assist plugins! Whether you're a Python developer, C++ enthusiast, or just getting started, our plugin system makes it easy to extend G-Assist's capabilities. Create custom commands, automate tasks, or build entirely new features - the possibilities are endless!
Each plugin lives in its own directory named after the plugin (this name is used to invoke the plugin):
plugins/
└── myplugin/ # Plugin directory name = invocation name
├── g-assist-plugin-my-plugin.exe # Executable
├── manifest.json # Plugin configuration
└── config.json # Settings & credentials
g-assist-plugin-<plugin-name>.exe- Executable file that executes plugin functionalitymanifest.json- Manifest file that contains:- name of the plugin executable
- plugin description
- list of functions and their parameters, with descriptions for each
tags- array of keywords used to describe the plugin's functionalitypersistent[true, false] - if the plugin should remain running throughout the entire G-Assist lifecycle
config.json- Config file containing any required information for the plugin (API key, usernames, other specifications) (⚠️ add to.gitignore)
💡 Tip: The plugin directory name is what users will type to invoke your plugin (e.g., "Hey myplugin, do something")
The manifest file acts as the bridge between G-Assist and your plugin. G-Assist automatically scans the plugin directory to discover available plugins.
-
Zero-Shot Function Calling
/fc What are the top upcoming games for 2025?The AI model automatically:
- Analyzes the user's intent
- Selects the most appropriate plugin
- Chooses the relevant function to execute
- Passes any required parameters
-
Direct Plugin Invocation
Hey logiled, change my keyboard lights to greenor
/logiled change my keyboard lights to green- User explicitly specifies the plugin
- AI model determines the appropriate function from the manifest
- Parameters are extracted from the natural language command
💡 Pro Tip: Direct plugin invocation is faster when you know exactly which plugin you need!
Explore our official plugins:
- Gemini AI Integration
- Logitech Peripheral Lighting
- Corsair Peripheral Lighting
- Nanoleaf Room Lighting
- Twitch Integration
- Spotify Music Player
- [More coming soon!]
Check out what others have built:
- [Your Plugin Here] - Submit your plugin using a pull request! We welcome contributions that:
- Follow our contribution guidelines
- Include proper documentation and examples
- Have been tested thoroughly
- Add unique value to the ecosystem
- Report issues on GitHub
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
We welcome contributions! Please see our Contributing Guide for details.