ASUS releases Zenni Claw beta for guided AI tasks
The public test build of ASUS Zenni Claw allows owners of compatible hardware to evaluate agentic artificial intelligence tools for daily planning and workplace tasks across both local hardware and...
ASUS has released a beta version of Zenni Claw, software designed to guide users through multi-step tasks on supported hardware using artificial intelligence.
Table Of Content
First introduced at Computex 2026, the application provides ready-made assistant skills for activities including work, travel and daily planning. The beta is available as a free download in select regions, subject to device compatibility.
Ready-made skills guide users through tasks
Zenni Claw is engineered to help users progress through an entire task rather than mainly respond to individual queries. ASUS describes the software as agentic AI, meaning it can take a single prompt to guide someone through a broader workflow. The beta is organised around four primary areas covering simpler setup, built-in skills for common scenarios, a choice between local and cloud-based AI processing, and safeguards such as guardrails and protection mechanisms.
Users can select a processing option based on their device, workflow and usage requirements, meaning the available experience may differ depending on the hardware and cloud service connectivity.
Local processing requires more capable hardware
The beta supports cloud AI options from Anthropic, ChatGPT and Google Gemini. However, availability may differ by region, beta version and service access.
For processing directly on the device, Zenni Claw relies on Gemma 4. ASUS notes that local performance and supported workflows depend heavily on the computer’s hardware, installed memory and available storage. The basic application requires a 64-bit installation of Windows 11 or later, at least 16GB of RAM and 20GB of available storage.
Running a local model raises the minimum memory requirement to 32GB, while users also need at least 20GB of storage alongside additional space for the software. Supported processor platforms for on-device execution include Intel Panther Lake H12Xe and AMD Strix Halo. For a higher-end local configuration, ASUS recommends an NVIDIA GeForce RTX 5090 graphics card with at least 24GB of video memory, alongside the requirement for 20GB of available storage and extra space for the software.
Availability varies by region and device
The Zenni Claw beta is available free for supported ASUS devices. Regional access and device compatibility may vary. ASUS is also actively seeking feedback during the beta period to refine the software, improve its existing assistant skills and add further capabilities in future versions.
