A Model Context Protocol (MCP) server that provides remote browser screenshots and vision analysis via USDC micropayments.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install screenshots-for-ai-agents
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install screenshots-for-ai-agents using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is screenshot-x402?
screenshot-x402 is a specialized skill for AI agents that enables remote browser rendering and visual analysis. Operating as an MCP server via Streamable HTTP, it allows agents to capture high-resolution screenshots of public websites or perform complex vision tasks, such as describing page content or extracting data, using the analyze_screenshot tool. This service is part of the emerging ecosystem of Openclaw Skills that prioritize secure, off-host processing for resource-intensive web tasks.
The skill stands out by integrating the x402 protocol, which facilitates seamless USDC micropayments directly between the AI agent and the service provider. By utilizing this infrastructure, developers can provide their agents with professional-grade web automation and vision capabilities without maintaining their own headless browser clusters, paying only for the resources they consume.
screenshot-x402 Use Cases
- Capturing high-resolution PNG or JPEG images of public web pages for documentation or archival.
- Performing visual data extraction where text-based scraping is insufficient or blocked.
- Integrating automated visual verification steps into AI-driven research workflows using Openclaw Skills.
- Offloading heavy browser rendering tasks from local agent environments to specialized remote infrastructure.
- Executing vision-based analysis on web content to answer specific natural language prompts.
How screenshot-x402 Works
- The AI agent connects to the screenshot-x402 MCP endpoint via a Streamable HTTP session.
- The agent initiates a tool call, such as take_screenshot, providing the target URL and desired viewport parameters.
- The server responds with a 402 Payment Required status, including metadata regarding the USDC cost and the target blockchain network (e.g., Base).
- The agent's x402-aware client signs a USDC authorization using a private key and submits the payment proof.
- Upon verification of the payment, the server renders the target URL on its remote infrastructure.
- The final result, containing the base64-encoded image and any requested vision analysis, is returned to the agent session.
screenshot-x402 Setup
To use screenshot-x402 within the ecosystem of Openclaw Skills, ensure you have an x402-compatible MCP client and a funded wallet.
- Configure your environment with your private key (do not share this with the LLM):
export X402_PRIVATE_KEY="your_private_key_here"
- Configure your MCP client to point to the remote endpoint:
# MCP Endpoint: https://screenshotx402.com/mcp
# Discovery: https://screenshotx402.com/discovery.json
- Use a client library like Cloudflare agents/x402 to automate the signing and retry logic.
screenshot-x402 Data Schema & Taxonomy
The skill returns structured data via the MCP standard. Below is the schema for successful tool executions:
| Field | Type | Description |
|---|---|---|
content |
Array | Contains the image data (base64) and text analysis. |
content[].type |
String | Either image or text. |
content[].data |
String | The base64 encoded string of the PNG/JPEG. |
_meta |
Object | Includes renderTimeMs and cached status. |
x402Network |
String | The blockchain network used for payments (e.g., Base). |
screenshot-x402 Advanced Features
- Smart caching: Use the
cacheTtlparameter to optimize costs by retrieving previously rendered screenshots. - Customizable viewports: Support for up to 4K resolution (3840x2160) and custom pixel ratios for retina-quality captures.
- Element manipulation: Use
hideSelectorsto remove intrusive elements like banners or popups via CSS selectors before capturing. - Vision integration: The analyze_screenshot tool combines rendering with LLM-based vision to provide text answers about the visual state of a page.
- Automated payment handling: Designed to work with Openclaw Skills and x402-aware clients for zero-intervention payment retries.
SKILL.md
Loading
Related Openclaw Skills
TokClaw Wallet
A shell-integrated AI agent skill for managing high-performance blockchain wallets on the TokClaw network.
Clawprint: LLC Formation for AI Agents
Clawprint enables AI agents to establish legal business identities, including Wyoming DAO LLC formation, EIN acquisition, and bank account setup under human sponsorship.
Pipeworx Jikan MCP
A Model Context Protocol server providing seamless, authentication-free access to anime and manga metadata from the Jikan v4 API.
Pipeworx IPInfo
A lightweight Model Context Protocol (MCP) server that provides instant IP address intelligence and location data without requiring an API key.
AI Video Editor Kaise Kare
An AI-powered video editing skill that automates trimming, transitions, and background music insertion for raw footage via cloud GPU rendering.
Free Video Generator Online AI
An AI-driven video generation skill that converts text descriptions and static images into professional 1080p MP4 videos via cloud GPU rendering.
Featured*
