Frequently Asked Questions

AgentVibe is a platform that empowers people to build and work with AI agents. Our first product is a free desktop app that lets you deploy and manage your own AI agents using open-source frameworks like OpenClaw and Hermes. Your agent can chat with you, remember things across conversations, send you morning briefings, and be reached through Telegram, Discord, WhatsApp, and other apps. Think of it as having your own personal AI assistant that lives on your machine.

Not at all. AgentVibe is designed for everyone. You pick a template, enter an API key, click Deploy, and your agent is running. Everything is done through the desktop app -- no terminal, no config files, no coding required.

That said, if you are technical, there is a CLI and full architecture docs for power users.

Yes. The desktop app is free, and running agents on your own computer is free. The only cost is the AI provider API usage (like paying Anthropic or OpenAI for model access). Most agents cost $5-15/month in API usage with moderate daily use.

Cloud deployments (for agents that need to be on 24/7) may require a commercial license and cloud hosting fees.

AgentVibe runs on macOS (Intel and Apple Silicon), Windows (x64 and ARM), and Linux (x64 and ARM). The desktop app is distributed as a DMG on macOS, an installer on Windows, and AppImage/DEB on Linux.

Yes. AgentVibe runs natively on Windows. You will need Docker Desktop installed (which is also free for personal use). The app installs like any other Windows program.

If your agent runs on your computer (the default), it pauses when your computer sleeps or Docker stops. When you open your laptop again, the agent automatically restarts and picks up where it left off. All memory and conversations are preserved.

If you need your agent available 24/7 (for example, a Telegram bot that should respond anytime), deploy it to the cloud instead.

Yes, Docker needs to be installed for local agent deployments. But you never need to interact with Docker directly -- AgentVibe handles all container management automatically. Just install Docker Desktop and forget about it.

The AgentVibe desktop app detects Docker and guides you through setup if it is not installed.

AgentVibe supports 22 AI providers including Anthropic (Claude), OpenAI (GPT), Google (Gemini), and many more. You can also use Ollama to run free, open-source models entirely on your own computer.

You can switch models on a running agent anytime -- no redeploying needed.

Yes. You can connect as many channels as you want. Your agent handles all of them simultaneously. Most people start with the desktop app for longer conversations and Telegram for quick messages on the go.

They do the same things. The desktop app is visual -- point and click, see your agents, chat in a nice interface. The CLI is for power users who prefer the terminal or want to automate things with scripts. Most people never need the CLI.

A template is a ready-to-use agent personality -- like an app in an app store. It comes with a name, personality, AI model, skills, and channel recommendations all pre-configured. Pick one and deploy. You can customize everything later.

Absolutely. Templates are starting points. After deploying, you can change the personality, switch the AI model, add or remove channels, install skills, and adjust scheduled tasks -- all from the desktop app.

Yes. You can deploy as many agents as you want. Each one runs independently with its own personality, model, channels, and memory. A personal assistant, a research agent, and a customer support bot can all run side by side.

The estimated costs in templates (like ~$8/mo) represent AI model API usage -- what you pay Anthropic, OpenAI, or your chosen provider. AgentVibe itself is free.

Running an agent on your computer has no infrastructure cost. Cloud hosting (if you need always-on availability) adds $3-20/month depending on the provider.

Everything stays on your machine. Agent conversations, memory, configuration, and your API keys are stored locally. AgentVibe does not send any data to our servers.

If you deploy to a cloud provider, your agent runs on that provider's infrastructure, but your API keys remain stored securely on your local machine.

On macOS, in the system Keychain (the same place Safari stores passwords). On Windows, in the Credential Manager. On Linux, in an encrypted local file. You can also provide keys via environment variables for automation setups.

Yes. The app is sandboxed, uses content security policies, and all internal communication is validated. Only explicitly allowlisted operations can be performed. See the architecture docs for technical details.

Make sure Docker Desktop is installed and running (look for the whale icon in your system tray). If it is running and still not detected:

1. Restart Docker Desktop
2. Open the AgentVibe app and check the Home page -- it shows Docker status
3. From the CLI, run agentvibe doctor for detailed diagnostics

First, check if the agent is running. Open the agent's detail page in the desktop app -- it shows the status. If stopped, click Restart.

If it is running but not responding to chat, the connection may need repair. The desktop app will show a connection status indicator with a Repair button if needed.

Large Docker image downloads can take a while on slower connections. If a deployment times out, you can increase the timeout:

export AGENTVIBE_STEP_TIMEOUT_MS=600000

Then retry the deployment. If you are deploying to a cloud provider, check that your API token is valid and has the right permissions.

In the desktop app, go to the agent's detail page and click the Delete button. This removes the container and all associated data.

From the CLI:

agentvibe destroy --id <deployment-id> --apply true --clean-data