OneClickClawDocs
$

Agent Control Panel

Last updated: March 2026

The Agent Control Panel is the most feature-rich page in your dashboard. This is where you configure everything about how your AI agent behaves: which AI provider and model it uses, what personality it has, which messaging channels it listens on, and how it responds to conversations. Every change you make here takes effect on your live agent.

Agent Control Panel — Main Sections

AI Provider Section

Provider dropdown with official brand logos (SVG) for 20+ providers. API key field with real-time validation. Model selector with 780+ models.

System Prompt Editor

Multi-line text area for agent personality and instructions. Save button applies changes immediately.

Channels Section

Connected channels list with status indicators. Red warning card if no channels configured. "Connect a chat channel" CTA button.

Agent Controls

Start / Stop / Restart buttons. Live agent status indicator (running, stopped, error).

AI Provider section

At the top of the Agent Control Panel, you find the AI Provider configuration. This is where you tell your agent which large language model to use and how to authenticate with the provider. The provider dropdown displays official brand logos (SVG) for each supported provider, making it easy to identify them at a glance.

Provider logos and selection

When you open the provider dropdown, you see official brand logos for over 20 supported providers, including OpenAI, Anthropic, Google, xAI, Mistral, DeepSeek, OpenRouter, Cohere, Meta, and more. Each logo appears next to the provider name, so you can quickly find the one you want.

  • Select a provider: Choose from any of the 23+ providers supported by OpenClaw. The dropdown shows all available options with their official logos.
  • Enter your API key: Paste the API key from your chosen provider. The key validates in real time to confirm it works before saving. Your key is encrypted at rest on your server.
  • Choose a model: Once a provider is selected, a second dropdown shows all available models for that provider. With 780+ models across 23 providers, you have extensive flexibility. Popular choices include Claude Sonnet for quality, GPT-4o for speed, and various open-source models through OpenRouter.

Note

You can switch providers and models at any time. Changes take effect immediately. There is no downtime when switching; the agent uses the new model starting from the next message it receives.

For a detailed walkthrough of provider setup with step-by-step instructions and API key links, see the Setting Up Your AI Provider article.

System Prompt configuration

The System Prompt section is where you define your agent's personality, behavior, and instructions. The system prompt is a block of text that gets sent to the AI model at the start of every conversation, shaping how the agent responds.

A good system prompt typically includes:

  • A role definition. Tell the agent who it is. For example: "You are a friendly customer support assistant for an online pet store."
  • Tone and style guidelines. Specify whether the agent should be formal, casual, humorous, or professional.
  • Topic boundaries. Define what the agent should and should not discuss. For example: "Only answer questions about our products and services. Politely decline off-topic requests."
  • Response format preferences. Tell the agent to use short answers, bullet points, step-by-step instructions, or whatever format suits your audience.

The system prompt editor supports multi-line text and holds several thousand characters. Changes are saved when you click "Save" and applied to new conversations immediately. For a deeper guide on writing effective system prompts with examples, see System Prompts & Personality.

Channel management

The Channels section lets you connect your agent to messaging platforms. OneClickClaw currently supports Telegram and Discord, with more platforms planned.

Channel connection warnings

If you have not connected any channels yet, the Channels section displays a red warning card with a "Not configured" badge. This card includes a prominent "Connect a chat channel" CTA button that takes you directly to the channel setup flow. Your agent cannot receive or respond to messages until at least one channel is connected.

  • Adding a channel: Click the CTA button or "Add Channel," select the platform (Telegram or Discord), and paste the bot token. The panel validates the token and shows a confirmation once the connection is live.
  • Removing a channel: Click the remove button next to any connected channel. The agent stops listening on that channel immediately. No data is lost; you can reconnect at any time.
  • Multiple channels: You can connect as many Telegram groups and Discord servers as you want. Your agent responds on all connected channels simultaneously using the same configuration and system prompt.

Tip

If you want different behavior on different channels (for example, a formal tone on your business Telegram and a casual tone on your Discord community), mention the platform in your system prompt with conditional instructions. For setup guides, see Telegram Setup and Discord Setup.

Agent controls

The agent controls let you manage the lifecycle of your OpenClaw agent process directly from the dashboard:

  • Start: Launches the agent process on your server. The agent begins listening on all connected channels.
  • Stop: Gracefully shuts down the agent. It stops responding on all channels. Your server remains running and accessible.
  • Restart: Stops and immediately restarts the agent. This is useful after changing configuration settings that require a full restart, or if the agent seems to be behaving unexpectedly.

The current agent status (running, stopped, or error) is displayed next to the controls so you always know the state of your agent.

Warning

Restarting the agent briefly disconnects it from all channels. Messages sent during the restart (typically a few seconds) may be delayed until the agent is back online.

Conversation history and logs

The Agent Control Panel includes a conversation history section that shows recent interactions your agent has had across all connected channels. This is invaluable for monitoring quality and troubleshooting issues.

Each log entry shows the channel, the user message, the agent response, the model used, and the timestamp. You can scroll through recent conversations to see how your agent is performing and whether it follows your system prompt instructions.

Note

Conversation logs are stored on your dedicated server. They are not sent to OneClickClaw's infrastructure. This means your conversations remain private and under your control. For more about data privacy, see Data & Privacy.

Tips for effective agent configuration

  • Start simple, iterate often. Begin with a short system prompt and refine it based on how the agent responds in real conversations. You can update the prompt at any time without downtime.
  • Test on a private channel first. Before deploying to a public group, connect a private Telegram chat or a test Discord server. Verify the agent's tone, accuracy, and behavior in a controlled environment.
  • Match the model to your needs. Larger models produce higher-quality responses but cost more per token. Smaller, faster models are better for high-volume, lower-complexity use cases. See Choosing the Right Model for a comparison table.
  • Review conversation logs regularly. Check how your agent handles edge cases, off-topic questions, and ambiguous inputs. Use these insights to improve your system prompt.
  • Use channel-specific instructions. If you connect to Telegram, WhatsApp, and Discord, note in your system prompt that the agent may be speaking to different audiences on each platform.