Note: The tool mentioned in this article, formerly known by various names including Clawdbot and Claudebot, has been rebranded to OpenClaw. This article has been updated to reflect the new name and installation instructions. For the latest information, please visit the official website: openclaw.ai.
A lot of questions have been circulating about OpenClaw (formerly Clawdbot) and how to leverage it with free AI models. There is, in fact, a powerful method to gain access to top-tier models like Claude Opus 4.5 and Gemini 3 Pro without cost, which is an incredible opportunity for developers and enthusiasts alike.
For those who have been following this publication, you may recall previous discussions about this tool. It’s a sophisticated messaging gateway that connects AI assistants like Claude Code and CodeX to various platforms, including Discord, WhatsApp, and even iMessage. This setup allows you to simply text your AI assistant to get tasks done. It’s a remarkably cool and efficient workflow.
The most significant question, however, always revolves around the models. Since OpenClaw is built upon the Claude Code foundation, it requires access to language models. What many don’t realize is that you can integrate it with Antigravity or Gemini CLI to tap into a suite of free, high-quality models. This article will show you exactly how to set up OpenClaw with Antigravity and Gemini CLI OAuth. The process is surprisingly straightforward.
The Power of Antigravity and Gemini CLI
Let’s break down what we’re accomplishing here. Antigravity, Google’s integrated development environment, offers a very generous free tier. This tier provides access to Claude Opus 4.5, Claude Sonnet 4.5, Gemini 3 Pro, and Gemini 3 Flash. The rate limits are quite decent for most use cases. The best part? OpenClaw includes built-in plugins that allow you to use these models seamlessly. Through the Antigravity OAuth plugin, you get complimentary access to this entire lineup.
The Gemini CLI OAuth option provides a similar pathway, connecting through the Gemini CLI authentication flow. Both methods work exceptionally well, and you can even combine them to achieve broader coverage and higher rate limits.
Step-by-Step Setup Guide
Here is a detailed walkthrough to get you up and running.
1. Install OpenClaw
First, ensure you have OpenClaw installed. If you haven’t installed it yet, you can do so with a simple command in your terminal.
curl -fsSL https://openclaw.ai/install.sh | bash
Once the installation is complete, run the onboarding command to start the setup wizard and get the gateway running on your system. The gateway is the core component that handles all messaging connections and routes them to the appropriate AI models.
openclaw onboard --install-daemon
2. Configure Antigravity OAuth
Now for the magic. To connect with Antigravity, you first need to enable the corresponding plugin.
openclaw plugins enable Google-Antigravity-O
This command activates the bundled plugin that manages the authentication process. Next, initiate the login flow.
openclaw models oauth login --provider Google-Antigravity --set-default
Simply follow the on-screen instructions to complete the connection. The plugin handles the entire OAuth flow, so you don’t need to manually paste any client IDs or secrets into your configuration file. It’s a truly hassle-free process.
3. Configure Gemini CLI OAuth
The process for the Gemini CLI is nearly identical. Start by enabling the plugin.
openclaw plugins enable Google-Gemini-CLI-O
Then, run the login command for the Gemini CLI provider.
openclaw models oauth login --provider Google-Gemini-CLI --set-default
Again, follow the on-screen prompts, and you’ll be set up in no time. The authentication tokens are stored securely on the gateway host via the CLI login flow, so once you complete the setup, you are ready to go.
Pro-Tip: Multiply Your Rate Limits
Here’s a powerful tip for advanced users. You can add multiple accounts to OpenClaw. This is a huge advantage because if you hit the rate limit on one account, OpenClaw will automatically rotate to the next available one. If you have two or three accounts, you effectively multiply your usage limits.
To add another account, just run the oauth login command again with a different account. It will be added to the rotation pool. This is especially useful for heavy coding sessions or if you have a team sharing the same OpenClaw instance.
Managing Your Models
To see all the models available through your configured providers, run the following command:
openclaw models list
This is a great way to verify that everything is working correctly. To set your default model, use the openclaw model set command with the <provider>/<model> format. For instance, to set the Antigravity Claude Opus model as your default, you would run:
openclaw model set Google-Antigravity/claude-opus-4.5
You can change this at any time. For quick, simple tasks, you might prefer the speed of Gemini Flash. For more complex coding or reasoning tasks, Claude Opus is the superior choice.
Expanding Your Options with Quen OAuth
If you’re looking for even more free options, Quen also provides complimentary model access through its device code OAuth flow. To set it up, enable the plugin first.
openclaw plugins enable Quen-Portal-O
Then, run the login command for the Quen provider.
openclaw models oauth login --provider Quen-Portal --set-default
With Quen, you gain access to their Coder and Vision models, which are also quite capable. The Vision model is particularly useful if you need your AI assistant to analyze screenshots or images.
Connecting Your Messaging Apps
Once your models are configured, the final step is to connect your messaging platforms. Run the channels login command, and it will walk you through connecting Discord, WhatsApp, Telegram, or any other supported platform.
openclaw channels login
The default gateway runs on localhost:18789. You can access a browser-based dashboard at that same address to monitor everything. From there, you can view active conversations, check usage statistics, and manage your connections all in one place.
Advanced Features
OpenClaw also comes with some powerful advanced features.
- Multi-Agent Routing: OpenClaw can automatically route different types of requests to different models. A simple question might be sent to Gemini Flash for a quick response, while a complex code-related query can be routed to Claude Opus. This helps you use your rate limits more efficiently.
- Custom Providers: If you have your own API access or want to use local models through a service like Ollama, you can configure that as well. OpenClaw autodetects if you have it running locally on the default port, allowing you to mix and match cloud and local models to fit your needs.
- Configuration File: For further customization, you can edit the configuration file located at
~/.config/openclaw/openclaw.json. Here, you can set up allowlists for users, configure group mention rules, and adjust a variety of other settings.
Summary
In summary, with a setup time of just five to ten minutes, you can have a fully functional AI assistant powered by elite models like Claude Opus 4.5 and Gemini 3 Pro, all for free. This assistant is accessible from your favorite messaging apps, creating a seamless and productive workflow.
This setup has proven to be incredibly reliable. While you might hit the free tier limits during extreme sessions, for normal daily use, you will likely not even notice them. The combination of OpenClaw’s flexibility and the generous free tiers from providers like Antigravity creates an unparalleled development tool.