Nous Research Brings Hermes Cloud's Desktop Agent Into Any Browser
Nous Research puts the full Hermes Desktop interface in a browser tab, letting cloud-hosted agents run 24/7 and stay reachable from any device.
- Hermes Cloud agents now open in a browser tab with the same UI as Hermes Desktop.
- Pick a name and model in Nous Portal to deploy a personal agent in roughly 60 seconds.
- Each agent runs 24/7 in its own container, handles scheduled jobs, and reaches you on any channel.
- Chats, group chats, plugins, and settings sync across phone, web, and desktop clients seamlessly.
- Plus ($20), Super ($100), and Ultra ($200) plans each include one Cloud agent out of the box.
- Compute is metered per second of uptime, with storage-only pricing when the agent is stopped.
Nous Research has added Hermes Desktop’s full interface to Hermes Cloud, giving hosted agents the same chats, group chats, plugins, and settings available in the desktop app. Hermes is Nous’s personal agent runtime for working with language models, tools, files, and messaging services. The cloud product already offered managed hosting; this release adds a browser interface that works across phones and computers.
One agent, two front ends
Developers create an agent in Nous Portal by choosing a name, model, and server size. Deployment takes about a minute, according to Nous. The resulting instance runs in a dedicated container with its own workspace and dashboard, while Nous selects the nearest region with available capacity.
The web app and Hermes Desktop can open the same hosted instance, so conversations, files, memory, and configuration remain attached to its cloud workspace. The local computer can shut down without interrupting scheduled jobs or connected messaging channels.
A persistent container behind the UI
| Capability | How it works | Practical use |
|---|---|---|
| Continuous availability | Nous describes the agent as available around the clock, with compute able to scale to zero while idle. | Tasks can continue when the developer’s computer is offline. |
| Persistent workspace | Files, memory, and configuration survive restarts and remain available across devices. | Long-running projects keep their state without manual synchronization. |
| Scheduled jobs | Natural-language schedules can trigger reports, backups, and briefings. | Recurring workflows run without an open browser or desktop session. |
| Messaging integrations | One agent can connect to Telegram, Discord, Slack, email, and the command line. | Each channel can access the same agent memory and workspace. |
| Parallel subagents | Subagents run in separate sandboxes with their own context histories. | Delegated tasks do not consume the main conversation’s token budget. |
The bill has three parts
Deployment requires an active subscription or at least $2 in account credits. Nous meters compute to the second while an agent is running and applies a lower storage-only rate when it is stopped. Paid subscriptions include hosting for one agent up to the plan’s listed size, while model inference and paid tool usage incur separate charges.
| Plan | Price | Included Cloud agent |
|---|---|---|
| Free | $0, usage billed separately | None |
| Plus | $20 per month | 4 GB RAM, 25 GB disk |
| Super | $100 per month | 4 GB RAM, 100 GB disk |
| Ultra | $200 per month | 8 GB RAM, 100 GB disk |
Nous currently operates regions in Amsterdam, Chicago, Dallas, Frankfurt, London, Los Angeles, Mumbai, Paris, San Jose, Singapore, Stockholm, Sydney, Tokyo, and Toronto. Placement is automatic and depends on proximity and available capacity.
One login, many model providers
Portal’s model catalog includes Anthropic Claude, OpenAI GPT, Google Gemini, Meta Llama, Mistral, DeepSeek, Qwen, xAI Grok, and Nous Research’s Hermes models. A single OAuth sign-in handles access across providers, removing the need to configure a separate API key for each one. Model inference remains a distinct part of the usage bill.
Chats, plugins, channel connections, and other cloud controls follow the desktop workflow. Developers configuring external services can consult the integration guide for setup details.
Jobs that outlive a laptop
- Research monitoring: An agent can collect sources overnight and prepare a report on a schedule.
- Development assistance: A coding conversation can continue from a phone, browser, or workstation while retaining the same files and history.
- Team automation: Scheduled pipelines can generate briefings and post them to Slack, Discord, or email.
- Parallel investigation: Subagents can inspect separate repositories or sources without filling the main agent’s context window.
For developers who would otherwise run Hermes on a virtual private server, Nous handles container provisioning, uptime, persistent storage, and browser access. Users still choose the model and server size, configure integrations, and account for inference and tool charges. The new interface makes that managed instance usable through the same workflow as Hermes Desktop.
Comments
