OpenClaw API Setup Tutorials

How OpenClaw setup works

Every OpenClaw assistant comes together from two choices: the LLM provider that powers its reasoning, and the channel where you talk to it. Pick a provider below for the model you want, then connect a channel — built-in chat works immediately, and Telegram, Slack, or WhatsApp let you reach the assistant from apps you already use.

Providers differ mainly in how you authenticate. Some, like OpenAI, Anthropic, Gemini, and MiniMax, support an OAuth or CLI login in addition to a plain API key, which can map to an existing subscription or coding plan. Others connect with a single API key. OpenRouter is the simplest way to reach many models through one key, and Ollama Cloud is the path for open-source models. Each guide below covers exactly where to get the credential and where to paste it in the dashboard.

Which provider should you choose?

If you already pay for ChatGPT or Claude, the OAuth and CLI-login paths for OpenAI and Anthropic let you reuse that subscription instead of buying separate API credits, so those are the easiest starting points for most people. MiniMax works the same way through its Coding Plan OAuth.

If you want flexibility, OpenRouter gives you one key that reaches dozens of models, which is useful when you are still comparing options or want to switch models without redoing setup. For privacy-sensitive or cost-controlled workloads, Ollama Cloud runs open-source models. Kimi and Z.ai connect with a single API key and are good picks if you specifically want Moonshot or GLM models. You can change providers later from the dashboard — the channel you connect (built-in chat, Telegram, Slack, or WhatsApp) stays the same.

Provider choice is also where most of your running cost is decided. If your bills are climbing, our guide on cutting OpenClaw API costs without chaos walks through trimming spend by picking the right model and auth path for each workload.

Why use managed OpenClaw hosting?

Running OpenClaw as your personal AI assistant requires careful setup of API credentials, messaging platform integration, and ongoing maintenance. OpenClaw Setup handles the infrastructure so you can focus on using your assistant.

• No server management: We handle gateway, updates, and monitoring for you the OpenClaw.
• Secure by default: Your API keys are encrypted and never exposed to your assistant runtime.
• Multi-platform ready: Run your assistant in built-in chat, Telegram, Slack, WhatsApp, or multiple channels simultaneously.
• Itemized hosting: Choose Light, Medium, or Large instances based on resource needs, and add Paperclip when you want the managed workspace. See pricing.

Telegram-specific setup

Need detailed guidance for Telegram? See our dedicated OpenClaw Telegram setup guide with bot creation, user ID retrieval, and configuration steps.

Need help getting started?

Once you have your API keys, head to the dashboard to connect your bot and launch your assistant.