Glossary of Terms

This glossary covers the key terms you encounter while using OneClickClaw. Terms are grouped alphabetically for quick reference. If you are new to AI agents and managed hosting, start here to build a solid foundation before diving into the rest of the documentation.

A through D

Agent

An AI-powered assistant that runs on your OneClickClaw server. Your agent is either OpenClaw or Hermes Agent, the one you picked for that server. It receives messages from connected channels (such as Telegram, WhatsApp, Discord or Slack), sends them to an AI model for processing, and delivers the response back to the channel. You define the agent's personality, knowledge, and behavior through its instructions. Each OneClickClaw instance runs exactly one agent, but that agent can serve multiple channels simultaneously.

API Key

A secret token issued by an AI provider (such as Anthropic or OpenAI) that authorizes your agent to use their models. Think of it as a password that lets your agent talk to the AI. Your API key lives on your own server. You generate it from your provider's dashboard. An API key is one way to connect your AI; a subscription sign-in is the other. See also: BYOK.

BYOK (Bring Your Own Key)

The pricing model used by OneClickClaw on every plan. You connect your own AI: an AI subscription you already pay for (such as a ChatGPT plan or SuperGrok) or an API key from an AI provider, and you pay that side directly. There is no built-in free model. OneClickClaw charges only for hosting infrastructure. This gives you full control over AI costs, provider choice, and model selection. There is no markup on AI usage. For a full explanation, see BYOK Explained.

Channel

A messaging platform connection where your agent listens and responds. Your agent offers many channels, among them Telegram, WhatsApp, Discord, Slack and Signal. You connect them inside your agent (on OpenClaw under Settings, then Channels; on Hermes Agent in the Hermes dashboard), which guides you through each one. You can connect multiple channels to a single instance, and your agent responds in all of them simultaneously using the same personality and configuration.

Dashboard

The web interface at oneclickclaw.io where you manage your OneClickClaw servers. From the dashboard you can view real-time server metrics (CPU, RAM, uptime), open your agent, view the Event Viewer, run diagnostics, see backups, choose when your agent upgrades, upgrade a paid plan below Pro, and monitor overall health. On Hermes Agent it also holds the AI Model, panel password and Webhooks cards. No terminal is needed for any of these tasks; SSH is optional.

Deploy / Deployment

The process of creating a new server and installing your agent (OpenClaw or Hermes Agent) on it. A deployment starts from the free trial or from a paid plan. OneClickClaw provisions a dedicated VPS, installs and configures all software, hardens security, sets up SSL certificates, and enables health monitoring. The entire process is automated and takes about 25 minutes for OpenClaw and five to seven minutes for Hermes Agent.

Diagnostics

A check of your server that you start with the Run Diagnostics button on the Troubleshooting card of your server page. It looks at your server and your agent and shows what it finds. If you contact support about an issue, running diagnostics first helps our team investigate.

E through I

Event Viewer

A real-time log display in your dashboard that shows everything happening on your server. Events include agent starts and stops, channel connections and disconnections, error messages, health check results, and update notifications. The Event Viewer uses a terminal- style display and helps you understand what your agent is doing at any given moment.

Gateway

The core component of OpenClaw that manages communication between your AI model and your messaging channels. The gateway runs on your server, listens for incoming messages from connected channels, forwards them to your AI provider, and delivers the response back. It operates via a WebSocket-based control plane on your server. You interact with the gateway through your agent's web interface, which your dashboard opens for you.

Health Check

An automated monitoring system that continuously verifies your server is running correctly. Health checks confirm that the server is reachable, OpenClaw is running, the gateway is responsive, and connected channels are active. If a health check fails, OneClickClaw's monitoring system triggers an alert and, in many cases, automatically attempts to recover the service.

Instance

A single OneClickClaw deployment consisting of one dedicated server running one agent, either OpenClaw or Hermes Agent. Each instance has its own guaranteed resources (CPU, RAM, storage), its own configuration, its own set of connected channels, and its own billing. Your subscription covers exactly one instance. If you need multiple independent agents, each requires its own instance and subscription.

J through N

Model Setup

The first screen OpenClaw shows when you open your agent for the first time. Press Connect provider and choose your AI: OpenAI, then Codex login (device code), for a ChatGPT plan (Plus or higher); xAI (Grok), then xAI OAuth, for SuperGrok or X Premium+; or a provider and its API key. Channels live under Settings, then Channels. On Hermes Agent, the AI is set from the AI choice card on the build page or the AI Model tab of your server page. For a full walkthrough, see Set Up Your Agent Inside OpenClaw.

O through P

OpenClaw

The open-source AI assistant gateway that powers every OneClickClaw deployment. OpenClaw connects AI models to messaging platforms, manages conversation history, handles session scoping, and provides a WebSocket-based control plane for real-time configuration. It supports 25+ AI providers and many messaging channels. OneClickClaw handles the installation, configuration and monitoring of OpenClaw on your behalf, and you decide when it moves to a newer version. You use it through its own web interface, which your dashboard opens for you.

Hermes Agent

The open-source AI agent made by Nous Research, the second agent you can pick for a OneClickClaw server. It connects AI models to messaging platforms and comes with its own dashboard. You choose its AI from the AI choice card on the build page or the AI Model tab of your server page, and connect channels in the Hermes dashboard. OneClickClaw installs, secures and monitors it on your behalf, and you decide when it moves to a newer version.

Provider

The company that supplies the AI model your agent uses. Supported providers include Anthropic (Claude), OpenAI (GPT), Google (Gemini), xAI (Grok), Mistral, DeepSeek, OpenRouter (hundreds of models through one key), and more. You connect a provider with your API key or, for some providers, by signing in with a subscription. On OpenClaw you choose your provider on the Model Setup page and can change it at any time under Settings, then Models, inside your agent. On Hermes Agent you choose it from the AI Model tab of your server page.

S through W

System Prompt

A set of instructions that defines your agent's personality, behavior, knowledge boundaries, and response style. The system prompt is sent to the AI model at the start of every conversation. It controls how your agent speaks, what topics it covers, how it handles edge cases, and what it refuses to discuss. A well-crafted system prompt is the most important factor in how useful and on-brand your agent feels to users. On OpenClaw you edit it in the Workspace Files card on your server page (the file SOUL.md).

Tier / Plan

The subscription level you choose when deploying. OneClickClaw offers three tiers: Starter (€14.99/month), Hobbyist (€25.99/month), and Pro (€49.99/month). Higher tiers provide more server resources (CPU, RAM, storage) and faster support response times. The same tiers and prices apply to OpenClaw and Hermes Agent. All tiers include the same core features: BYOK, agent upgrades whenever you choose, health monitoring, SSL, and unlimited channels. You can try a real Starter server free for 7 days first, with no card. See Choosing a Plan for a full comparison.

VPS (Virtual Private Server)

A dedicated virtual machine that runs your agent (OpenClaw or Hermes Agent). Each OneClickClaw instance gets its own VPS with guaranteed resources (not shared with other customers). This means your agent's performance is consistent and not affected by what other users are doing. All VPS instances are hosted in the EU (Denmark) on AMD EPYC processors with DDR4 RAM and NVMe SSDs.

Webhook

A mechanism where one system sends an automated notification to another when a specific event occurs. OneClickClaw uses webhooks internally (for example, Stripe sends a webhook when you complete payment, which triggers server provisioning). These run behind the scenes and need nothing from you. Hermes Agent servers also have a Webhooks card on the server page for your own use.

Free Trial

Seven days on a real Starter server, free, with no credit card: sign in with Google and start it from your dashboard, for OpenClaw or Hermes Agent. When the 7 days end the trial server is paused and kept with its data for 3 more days, then deleted, unless you pay for Starter before then. Paying keeps the same server, and you can move up to a larger plan afterwards.