Managed Hermes Agent
Not installed yet? On Rollin it is 1 click — with human support.
- 1-click deployment + free subdomain
- Native migration from Open Claw
- Support 24/7 on WhatsApp and Discord
Installed it (or about to)? This is the path of the first 30 minutes — connect, choose the brain and teach it your style.
First steps with Hermes Agent: (1) connect WhatsApp via the QR Code in the panel; (2) choose the model — local (free, private) for day-to-day, API (GPT/Claude) for extra power; (3) introduce yourself — who you are, what you do, how you like your answers: memory and the learning loop start from there; (4) enable skills (agentskills.io) as needed. On Rollin Host, installation is 1 click and support is human, on WhatsApp.
With Hermes live (on Rollin, that is 1 click), the first action is connecting a channel. Of the 22 supported, WhatsApp is the standard: open the channels area in the panel, scan the QR Code with the number that will belong to the assistant and test with a "hi".
Then, the brain: Hermes works with 200+ models. Start with a local one (Llama 3 and similar) — zero per-message cost, data inside the server. If a task demands more (long analyses, code), plug in a model via API with your key. Switching later is free.
Hermes improves with use — but you speed it up a lot in the first days by being explicit: say who you are, your projects, and the format you prefer ("short answers", "always in bullet points", "ask me before assuming"). Correct it freely: every correction feeds the learning loop.
When you miss a capability, look for a skill (compatible with agentskills.io) — the way to add abilities without touching code. And remember: any friction, support on WhatsApp solves it with you.
Not installed yet? On Rollin it is 1 click — with human support.
Three steps: (1) connect a messaging channel — almost always WhatsApp, scanning the QR Code in the panel; (2) choose the AI model (a local one, like Llama 3, or one via API with your key); (3) introduce yourself: tell it who you are, what you do and how you like your answers. Memory starts working from there.
Rule of thumb: start with the key of the provider you already use (GPT, Claude, Gemini) and switch models whenever you want, without losing memory. Hermes accepts 200+ models — and you can switch anytime.
By use. The learning loop observes interactions: what you correct, the format you ask for, the recurring topics. Speed it up in the first days by saying explicitly how you prefer your answers (short? with sources? in bullet points?).
Extra abilities you add to the agent — Hermes is compatible with agentskills.io. It is like installing apps on the assistant: new capabilities without touching code.
On Rollin Host WhatsApp — human support, no ticket queue. We join you in the panel if needed. There is also the Discord community.
Get started in 5 minutes. Free migration (on the semiannual plan), 24/7 support in Portuguese and a 7-day money-back guarantee (30 days on website and WordPress hosting).
We use cookies to analyze traffic, improve your experience and personalize content. You decide what to accept — see our Cookie Policy.
Choose which categories you allow. Necessary cookies are essential for the site to work and cannot be turned off.
Essential for navigation, security and basic site operation. They do not track you.
Help us understand, anonymously, how visitors use the site (Google Analytics).
Let us measure campaign effectiveness and show relevant ads (Meta Pixel).