OpenClaw integrates with PhanthyCloud! Paradigm launches PhanClaw: Building secure, controllable, and low-cost Agent infrastructure

As the most sought-after open-source Agent framework globally, OpenClaw has quickly become the top choice for developers building digital assistants thanks to its exceptional task decomposition and complex environment interaction capabilities. It not only redefines human-computer interaction but also triggers a paradigm shift in the global tech community from 'conversation' to 'full automation'.

Today, FanShi announced that OpenClaw has officially integrated with PhanthyCloud, simultaneously launching PhanClaw, an Agent platform deeply embedded within PhanthyCloud. As a powerful extension of the OpenClaw ecosystem, PhanClaw aims to provide users with secure, controlled, and low-cost 'turnkey' clusters of digital assistants.

What exactly is OpenClaw?

OpenClaw is the most popular AI action-oriented tool globally in 2026. It can operate a computer like a human, 'understanding' buttons, input fields, and menus on the screen while simulating mouse clicks, keyboard inputs, and file transfers. You just need to give it a command (like 'book a ticket for me' or 'search for information and create a PowerPoint'), and it will open the browser, search for information, fill out forms, and complete the task independently. Due to its logo featuring a striking red lobster, symbolizing that AI has developed strong 'claws' to work with, it is affectionately referred to as 'Lobster'.

Currently, the popularity of "Lobster" is beyond imagination: it surpassed 100,000 stars on GitHub in just 10 days. Both geeks and corporate teams, domestically and internationally, are aggressively deploying this tool. This craze has even spilled over into the hardware market—due to the large number of users needing standalone devices to run OpenClaw around the clock, Mac minis have been repeatedly sold out worldwide.

However, behind this powerful 'lobster', there are thorny issues that trouble developers. First, it's expensive. In traditional models, merely maintaining operational startup costs could reach over $1,000 per day, plus costly hardware depreciation and maintenance expenses, deterring many teams. Second, there's risk; OpenClaw needs access to large amounts of local information during operations. Without effective protection, core data such as personal privacy and account passwords can easily be exposed to risks. To address these pain points, FanShi provides a one-click deployment architecture compatible across all scenarios, optimizing the usage environment for OpenClaw.

Extremely low usage cost: Leveraging PhanthyCloud’s resource pooling technology, PhanClaw effectively addresses the high daily hardware depreciation and maintenance costs associated with OpenClaw’s traditional model, significantly reducing computing power costs and avoiding 'high consumption from the moment of startup'.

Persistent data: PhanClaw promises not to clear user data proactively. By creating a secure operational sandbox, PhanClaw gives Agents 'long-term memory'. Users no longer need to retrain each time they start up, ensuring workflow continuity.

Data security: Traditional cloud-based large model API modes may expose all local information—personal privacy, social media, banking accounts, passwords—to external vendors uncontrollably. FanShi, through PhanthyCloud-powered private ModelHub, supports launching large model services directly in an intranet private cloud environment, ensuring no information leaves the premises, thus eliminating privacy leakage risks at the source.

Rich model matrix: PhanClaw offers users a highly competitive model matrix via ModelHub, including top-tier models like GLM-5, MiniMax M2.5, Kimi K2.5, which rival Claude in performance, as well as multiple other cost-effective, fast-response models. Additionally, through deep integration with PhanthyCloud, PhanClaw allows one-click access to world-class model capabilities across various domains in the Phanthy Pantheon, enabling more complex task scheduling and execution.

PhanClaw is not isolated but rather a core component nested within FanShi's technical stack, with a clearly structured internal architectural logic.

Application Layer (Agent): As the entry point for direct user interaction, PhanClaw achieves compatibility and enhancement of OpenClaw's native capabilities.

Model and Engine Layer (ModelHub + Phanthy Pantheon): PhanClaw can invoke models such as GLM-5, MinimaxM2.5, KimiK2.5 in real-time, and seamlessly integrate with the world models from Phanthy Pantheon to address more professionally challenging tasks.

Scheduling Layer (HAMi vGPU): Through the HAMi scheduling engine, PhanClaw implements fine-grained partitioning and isolation of underlying GPUs, ensuring maximum stability even during multi-instance operations.

Base Layer (PhanthyCloud + Hardware Infrastructure): PhanthyCloud provides full environment adaptation from public to private cloud, ensuring each computation runs on a solid foundation.

Starting today, PhanClaw is officially launched on PhanthyCloud. Paradigm is committed to enabling every developer and professional team to find their own 'all-purpose assistant' in PhanClaw, evolving from simple 'conversations' to efficient 'execution'.

Product consultation link: http://cloud.phanthy.com

Paradigm Group $PHANCY (06682.HK)$ is a global leader in general artificial intelligence technology, with the mission of 'AI for everyone', dedicated to empowering various industries through the technical approach of 'AI agent + world model'. The company was founded in 2014 and became a group in 2025, encompassing business segments such as enterprise services (Fourth Paradigm), large models and AGI (Pantheon), consumer electronics (Paradigm Pilot), smart energy (Paradigm Ark), and smart sports (Paradigm NetMotion). To date, the company has successfully implemented over 10,000 AI applications in finance, retail, healthcare, and other fields, consistently promoting AI accessibility and helping enterprises achieve sustainable growth.

38K Views