🦞 OpenClaw Assistant#
The OpenClaw Assistant (Lobster Assistant) is an AI-powered Q&A assistant built into the FluxVLA documentation site. You can ask it questions at any time, and it will generate answers based on the documentation content, helping you quickly locate the information you need.
How to Open#
In the upper-right corner of the documentation webpage, you will find a dark blue “🦞 OpenClaw Assistant” button. Click it to open the conversation window.
Tip
This button appears on every page of the documentation, allowing you to invoke the assistant from any page at any time.
Conversation Window Overview#
Upon opening, you will see a pop-up conversation panel containing the following areas:
Title bar: Displays “FluxVLA Documentation Assistant”, with a close button (×) in the upper-right corner.
Chat area: Shows the conversation history between you and the assistant. Upon first opening, a welcome message and recommended questions are displayed.
Input field: Enter your question at the bottom and click the “Send” button or press Enter to submit.
Footer notice: A reminder that “Responses are generated by AI and are provided for reference only. Please verify against the actual documentation and source code.”
Three Response Modes#
On the left side of the input field, there is a mode selection badge. Click it to switch between three modes:
Mode |
Description |
Estimated Time |
|---|---|---|
⚡ Quick Answer |
For everyday questions with rapid responses, suitable for simple and straightforward queries |
≈2-3s |
⚖️ Balanced Mode |
Strikes a balance between speed and detail, suitable for questions requiring a certain degree of context |
≈30s |
🧠 In-Depth Analysis |
Suitable for in-depth reasoning and explanation of complex questions, with more comprehensive document retrieval and synthesis |
≈120s |
Recommended Questions#
Upon first opening the conversation window, the assistant will display several recommended questions for quick selection:
What is FluxVLA?
How do I get started with FluxVLA?
What examples are available in the documentation?
What precautions should be taken when using FluxVLA?
You may also enter any question directly in the input field.
Usage Tips#
Be specific: Instead of asking “How do I train?”, asking “How do I perform multi-node distributed training of OpenVLA using FSDP?” will yield more precise answers.
Leverage the current page: The assistant is aware of the page you are currently viewing. You can directly ask “What does xxx mentioned on this page mean?”.
Multi-turn conversations: You can ask follow-up questions within the same session, and the assistant will retain context.
Verify responses: AI-generated answers are provided for reference only. For critical configurations and code, please refer to the original documentation and source code.