Holo3.1
Holo3.1 is H Company’s family of fast and local computer-use agents, built to operate across web, desktop, and mobile environments while integrating more smoothly into different agent frameworks and deployment targets. Based on the Qwen family, Holo3.1 improves robustness across the environments where computer-use agents are actually deployed, addressing the distribution shifts that appear across mobile devices, alternative agent harnesses, and different execution frameworks. The release expands Holo3’s capabilities beyond browser and desktop control, with major gains in mobile automation, including AndroidWorld improvements from 67% to 79.3% for the 35B-A3B model and from 58% to 71% for the smaller 4B and 9B variants. Holo3.1 also introduces native support for function-calling protocols in addition to structured JSON outputs, helping teams deploy the model inside third-party agent stacks with near-parity between function-calling and native execution.
Learn more
Holo AI
Organize your thoughts into incredible compositions with a few clicks. It's built for anyone writing anything. Features aimed at letting you explore, unrestrained. Novels, short stories, and fanfiction, our metadata UI lets you tune the AI to evoke from a myriad of different fandoms, genres, and authors. Our prompt tuning capabilities let you train our model on the custom data that you provide. This can be as simple as feeding your AI purely Edgar Allan Poe or as complicated as designing a chatbot with transcript data. Configure Holo AI to read generations to you out loud and can choose from 6 different AI voices. HoloAI stories and generation metadata (like key-context pairs) are client-side encrypted. That means the devs have no technical way to access them or give them to anybody else. Datasets for every type of work and end-to-end encryption.
Learn more
Qwen2
Qwen2 is the large language model series developed by Qwen team, Alibaba Cloud.
Qwen2 is a series of large language models developed by the Qwen team at Alibaba Cloud. It includes both base language models and instruction-tuned models, ranging from 0.5 billion to 72 billion parameters, and features both dense models and a Mixture-of-Experts model. The Qwen2 series is designed to surpass most previous open-weight models, including its predecessor Qwen1.5, and to compete with proprietary models across a broad spectrum of benchmarks in language understanding, generation, multilingual capabilities, coding, mathematics, and reasoning.
Learn more
Surfer H
Surfer H from H Company is an autonomous web-agent platform built to understand and navigate user interfaces like a human by combining three modular models; a policy model that plans tasks, a localizer model that identifies UI elements visually, and a validator model that checks outcomes. The agent works purely through the browser interface with no special API hooks, enabling it to scroll, click, type, and complete real-web tasks such as booking hotels, comparing product deals, or extracting structured information. When paired with H Company’s open-weight vision-language models, Surfer H achieved state-of-the-art performance on the WebVoyager benchmark (92.2% accuracy at around $0.13 per task) and supports deployment locally, via Docker, or on cloud infrastructure. Use cases span web automation, QA testing without brittle scripts, data harvesting, and intelligent workflow agents that interact with the web directly as a human would.
Learn more
