RAZER AIKIT
START LOCAL, SCALE TO CLUSTERS BUILD ADVANCED AI END-TO-END
Experience the full power of Razer’s GPU-powered devices with Razer AIKit, the open-source AI development toolkit built for engineers and researchers. Designed for easy, out-of-the-box setup, Razer AIKit delivers cloud-grade GPU acceleration and scalability for AI developers to run LLMs locally over a single or cluster of GPUs.
Features
Run and fine-tune over 300,000 LLMs locally with familiar frameworks, achieving full performance and AI privacy without cloud costs.
When local power isn’t enough, connect your GPUs into a single cluster for seamless scaling and optimal performance with minimal setup.
Fully open-source, Razer AIKit invites developers to customize and contribute, advancing high-performance local AI development together.
Case Studies
Razer blade 18 x openclaw
YOUR OWN AI ASSISTANT.
RUNNING LOCALLY. NO CLOUD.
We put Razer AIKit to work on a single machine. On a Razer Blade 18, AIKit handles all model inference locally — running OpenClaw as a real-time personal AI assistant that monitors tasks and automates workflows in the background. No cloud dependency. No latency overhead. Data stays on your machine, and setup takes minutes.
On-device inference. 50+ automations. Zero cloud calls.
That’s what AIKit looks like when you bring it home. Try it now below.
Razer x akash
razer aikit at scale.
proven in production.
We put Razer AIKit through a real test. During the AVA Mini campaign, AIKit ran on a globally distributed pool of consumer GPUs via Akash Network’s decentralised compute marketplace — serving thousands of personalised AI-generated images over four days. Zero outages. Zero manual intervention.
30 images per minute. 3.24 seconds per image. $0.01 per image.
That’s what AIKit looks like when you take it beyond a single machine.
DEVELOPED WITH RAZER AI HARDWARE
POWERING THE NEXT AI BREAKTHROUGH
Built on more than 20 years of high-performance engineering, the Razer Forge AI Dev Workstation is our flagship system for professional AI development. Purpose-built to pair seamlessly with Razer AIKit, it supports demanding training, inference, and simulation workloads with robust multi-GPU scaling, large memory capacity, and high-speed interconnects.
Start building with Razer AI hardware and the open-source Razer AIKit to move from AI development to production—without compromise.
FAQ
What GPU do I need for model inferencing and fine-tuning?
Razer AIKit supports all NVIDIA GPUs (Workstation/Consumer, Data Center, and Jetson) with compute capability 7.0 or higher. See the full list on NVIDIA’s GPU page.
What tools and frameworks does Razer AIKit include?
Razer AIKit is built on vLLM, Ray, and makes use of LlamaFactory for fast inference, efficient fine-tuning, and multi-GPU scaling.
What models are supported?
Any vLLM-compatible model from Hugging Face Hub – over 300,000 LLMs available and ready to run locally.
Is Razer AIKit only for Razer devices?
Razer AIKit is optimized and tested on Razer devices for peak stability and performance. However, it’s not limited to Razer hardware – Razer AIKit runs on any system equipped with a compatible GPU.
How can I use a running model?
Razer AIKit comes with OpenAI API integration, enabling connections to tools like Open WebUI, Continue Coding Assistant, AnythingLLM, and AI Dev Gallery.
Use them for tasks such as code generation, QA companions, research assistance, and local AI experiments.
