Build AI Tools and Win Big with NVIDIA's Project G-Assist Plug-In Hackathon

Project G-Assist is a powerful new AI assistant, available via the NVIDIA App, for GeForce RTX systems. Running locally on RTX hardware, the tool is easily accessible via the NVIDIA Overlay at any time, with an interface reminiscent of popular AI tools like ChatGPT that accepts both voice and text commands.

The difference here is that Project G-Assist runs locally, powered by NVIDIA's groundbreaking AI hardware on its GeForce RTX GPUs, so it can do everything from provide real-time information about a game's performance (with charts) to offering recommendations on how to improve efficiency, hit 100+ FPS in a competitive shooter like Marvel Rivals, overclock your GPU, and more.

As a Small Language Model (SLM) with 8 billion parameters tuned for GeForce RTX systems and gamers, you can communicate your questions, commands, or thoughts with natural language, which will then call on a range of NVIDIA or third-party APIs to execute these commands in real-time.

From something simple, like "What are my CPU and GPU specs?" to "Can you optimize the graphics in Cyberpunk 2077 so I can hit 60 FPS with ray-tracing?" things really begin to open up once plug-ins enter the mix. Project G-Assist was designed with the PC gaming community in mind, with NVIDIA offering a wide range of resources and tools on GitHub for users to create their own G-Assist plug-ins.

From experienced Python and C++ coders looking for a rapid way to access applications and third-party tools, there's even support for beginners looking to develop their own AI tool to suit their needs. From accessing smart devices with IFTTT, firing up an app, queing up the next song on Spotify, NVIDIA's GPT Plug-in Builder allows you to transform your ideas into reality and extend Project G-Assist's capabilities into fascinating new areas.

6

And there's no time like the present because NVIDIA's Plug and Play: Project G-Assist Plug-In Hackathon is currently running from now until Wednesday, July 16. This event allows users to submit their custom G-Assist plug-ins for a chance to win RTX Blackwell gear, including a powerful GeForce RTX 5090 Laptop, a GeForce RTX 5080 Desktop GPU, and the opportunity to meet the G-Assist Team.

Some cool examples of current Project G-Assist plug-ins

The sky, as they say, is the limit to what you can achieve with a custom Project G-Assist plug-in. However, you might be wondering what's possible - and with that, let's go through some of the most popular Project G-Assist plug-ins currently available.

Project G-Assist runs locally on an optimized Small Language Model (SLM), but that doesn't mean you're unable to access a powerful Large Language Model (LLM) via the G-Assist interface. The Google Gemini plug-in enables users to access Gemini's powerful search-based queries and capabilities in real-time without the need to open an additional window.

6

NVIDIA's Discord plug-in for G-Assist adds integration with one of the most popular chat and community apps and platforms for gamers. This powerful plug-in enables you to send text messages, performance charts, and even the ability to send game clips recorded via Shadowplay to a Discord channel.

The G-Assist Spotify plug-in is another example of a powerful way to interact with a popular app via the G-Assist interface. With voice commands or text, you can play your favorite music, control playback, and even manage playlists. Again, without the need to open an additional window.

6

IFTTT, or If This Then That, provides access to IFTTT applets and smart devices, offering direct control over features such as lighting and other elements. You can even access RSS feeds to get all of the latest PC hardware and gaming news from TweakTown while taking a break from the action in an FPS lobby or a town in an RPG.

The final plug-in example, Twitch, allows you to check any streamer to see if they're online, what they're playing, and the current view count.

How to create a Project G-Assist plugin with no programming knowledge using GPT

You don't have to be a coder to jump straight into plug-in development, as the G-Assist Plug-In Builder is a ChatGPT-based app that is all about automating the coding process while creating all of the files and tools you need to make your plug-in dream a reality. With natural language, no coding, or limited coding knowledge, you can be up and running in minutes. If you're wondering how it all works, here's a breakdown of the process of creating the Twitch plug-in with the G-Assist Plug-In Builder.

6

1. Using the G-Assist Plug-In Builder, use natural language to tell it what plug-in you're looking to create. For the Twitch plug-in, this would be something like, "I want to create a G-Assist plug-in that interacts with Twitch to see if a user is live or not."
2. The Plug-In Builder will then offer an easy-to-follow overview of what's required and the files it will generate. It will also inform you that it requires a Twitch Client ID and token (Secret), along with instructions on how to obtain them.
3. In this specific example, the Plug-In Builder even notes that Twitch tokens can expire, so it offers a solution for resolving this issue while updating the relevant code. As the plug-in directory is admin-restricted, all sensitive credential information is kept secure.
4. From there, it's as simple as telling the Plug-In Builder to create the Twitch plug-in, which will then generate the relevant Python script alongside the manifest, requirements, and config files.
5. These can be copied over to the plug-in template available on GitHub, where you can review the code and make changes. From there, it's as simple as generating the plugin and copying it over to the G-Assist plug-ins directory, with the folder name being what you'd say when calling on the new AI-powered tool.

Project G-Assist Plug-In Hackathon - Start creating and win big

NVIDIA is currently looking to highlight and showcase innovative and creative Project G-Assist plug-ins that are easy to use. For those looking to enter NVIDIA's Plug and Play: Project G-Assist Plug-In Hackathon, simply register here, get creating, and then share what you've cooked up with the #AIonRTXHackathon tag on social media. Submit your project via this form by Wednesday, July 16, for your chance to win some RTX goodness.

6

What's up for grabs? Well, winners will be selected on Wednesday, August 20, with the first-place winner scoring a GeForce RTX 5090 laptop, second place a GeForce RTX 5080 GPU, and third place a GeForce RTX 5070 GPU.

Project G-Assist is all about the PC gaming and AI community, so be sure to join the NVIDIA Developer Discord channel to collaborate, share your ideas, and even get the support you need to make them a reality. NVIDIA's GitHub repository features everything you need to get started on your G-Assist journey, with samples to download, step-by-step instructions, and detailed documentation covering all aspects of Project G-Assist.