Nous Research officially released Hermes Desktop v0.15.2 public preview on June 2, under the MIT open-source license. The desktop application supports macOS 12+, Windows 10/11, and Linux, integrating persistent memory, natural language scheduling, over 300 models, and multi-platform real-time messaging—eliminating the need for terminal-based operation.
From Terminal to Desktop: A UI Revolution for Hermes
Hermes Agent is an open-source autonomous AI agent built by Nous Research, distinguished by its self-learning capability. Once it masters a new task, it saves the method as a reusable skill file, so it doesn't start from scratch next time. But the barrier was always the interface: users had to launch a terminal, run curl commands, and interact line by line. For non-technical users, that's a high wall.
Its main open-source rival, OpenClaw, had a graphical interface from the start, which helped it build a non-technical user base early on. Hermes boasted a more compelling self-learning skill loop, but the learning curve was steeper. Hermes Desktop now closes that gap.
Full Suite of Official Desktop Features
Built with Electron and React on a Python backend, Hermes Desktop shares the same core agent engine as the CLI version—same persistent memory, same skill system, same configuration. No work gets left behind.
The feature set includes: persistent memory for storing projects and solutions, natural language scheduling for recurring tasks and reports, built-in web browsing and image generation, and over 300 models accessible through Nous Portal. Users can delegate tasks to sub-agents (smaller Hermes instances running isolated jobs), and interact with the same agent across Telegram, Discord, Slack, WhatsApp, Signal, and email. Sandbox environments include local, Docker, SSH, Singularity, and Modal.
MIT Open Source and Free Tier
The application is released under the MIT license: anyone can review the source code, self-host, or modify it without permission. Nous Portal offers a free usage quota, with paid plans (Plus, Super, Ultra) providing monthly credits and access to additional models.
Nous Research notes that Hermes Desktop is still in public preview, so rough edges are expected. The GitHub repository is live, and the team is actively collecting user feedback. With the official desktop launch, the open-source AI agent market competition has officially moved from the CLI domain to the desktop arena.