Razer AIKit: requirements, official setup, and its relationship with AVA

A source-checked overview of Razer’s local-AI toolkit. This is not a substitute for Razer documentation or a hands-on installation guide.

A local-AI toolkit, not an AVA-specific SDK

Razer introduced AIKit at CES 2026 as an open-source development toolkit for running, tuning, and scaling AI workloads locally. It is a general developer tool; Razer has not presented it as a public SDK for modifying Razer AVA.

Razer introduced the product at CES 2026 .

What the official project documents

The official repository documents AIKit CLI, vLLM inference, Ray for distributed workloads, and LlamaFactory for parameter-efficient fine-tuning. It also publishes the Docker Quick Start and compatibility information.

Technical capabilities and commands are maintained in the official AIKit repository .

Local inference

Run compatible AI workloads on your own NVIDIA hardware.

Scaling with Ray

Use distributed computing when a workload needs more than one GPU or node.

Open-source project

The repository is published under Apache License 2.0.

Evolving scope

Razer announced image, video, audio, and Arm64 expansion on April 30, 2026.

What is confirmed about AVA

AIKit and Razer AVA belong to the same Razer AI ecosystem, but public material does not define AIKit as the AVA SDK or document a permanent integration. Razer said AIKit powered the AVA Mini image-generation campaign; that is evidence of a specific use case, not proof of a permanent product architecture.

Status checked on September 23, 2026

Razer’s April 30 update documents the omni-modal expansion and the AVA Mini use case. It does not announce AIKit as an AVA modding SDK. Read the source .

Requirements documented by the project

NVIDIA GPU with Compute Capability 7.0 or higher.
Windows 11 through WSL2, or Ubuntu 22.04 / 24.04.
Docker Engine and NVIDIA Container Toolkit.
A current NVIDIA driver; the project recommends the Studio Driver on Windows.
Razer hardware is not listed as a requirement.

Always check the current official compatibility guidance before starting.

Official Docker Quick Start

Source, not a claimed test

The following commands come from Razer’s official README. Review the current source and your environment before using them.

1

mkdir -p $HOME/.cache/huggingface

2

docker run -it --restart=unless-stopped --gpus all --ipc host --network host --mount type=bind,source=$HOME/.cache/huggingface,target=/var/aikit/.cache/huggingface --env HUGGING_FACE_HUB_TOKEN=<YOUR_TOKEN> razerofficial/aikit:latest

3

jupyter lab --ip="0.0.0.0"

Before using AIKit on Linux

The official project lists Ubuntu 22.04 and 24.04. Check the current README for driver, Docker, and NVIDIA Container Toolkit instructions before running a container. AvaSDK has not independently benchmarked or tested this procedure, so this page deliberately does not claim installation results.

Follow verified changes

Subscribe for updates when Razer publishes a material change to AIKit or AVA.

Subscribe

Frequently asked questions

Is Razer AIKit free?

The official repository is public and licensed under Apache License 2.0.

Do I need a Razer device?

No. The official requirements describe compatible NVIDIA systems, not Razer-only hardware.

Is AIKit the Razer AVA SDK?

No public Razer source reviewed here defines it as an AVA-specific SDK. See the AVA status page for the currently documented developer paths.

Primary sources