Fetch Rich Metadata from URLs
Pull the title, description, thumbnail image, and Open Graph data from any public URL.
Populate Link Cards and Embeds
Fetch metadata from shared URLs to power link card previews in your application or content feed.
Quick answer: Use the Link Preview tool through ToolRouter to populate link cards and embeds directly from Claude, ChatGPT, Microsoft Copilot, and OpenClaw — connect once, then drive it with plain-language prompts. No code required.
ToolLink cards — the rich previews that appear when a URL is shared in a messaging app, a knowledge base, or a social feed — require fetching and parsing metadata from the target URL. Building this as a custom feature means writing an HTTP client, parsing HTML, extracting the right meta tags, and handling edge cases for the thousands of sites users might share.
The preview_url skill does this reliably for any public URL, returning a structured object with all the fields needed to render a link card: title, description, thumbnail image, site name, and URL. No custom parsing required.
Developers building Slack-like messaging apps, teams maintaining curated resource libraries, and product teams adding link sharing to their platform all use this to add rich link previews without building a custom unfurl service.
How to populate link cards and embeds with Claude, ChatGPT, Microsoft Copilot, and OpenClaw
Claude fetches link card metadata and can immediately format it into the structure your application needs. Ask it to return the data as a JSON object matching your schema, or describe the card layout and it formats the output accordingly.
Connect ToolRouter to Claude
1Open connector settings Open Settings
2Add a custom connector with these details
Name
ToolRouterURL
https://api.toolrouter.com/mcp3Let Claude set you up Open Claude
How to populate link cards and embeds with Claude
Once connected (see setup above), use the Link Preview tool:
- Ask: "Get link card metadata for: [paste URL]"
- Claude returns the structured preview data
- Ask: "Format this as a JSON object with title, description, image, and url fields"
- Use the formatted data to render your link card component
Example prompt for Claude
Try this with Claude using the Link Preview tool
Fetch link preview data for this URL and return it as a JSON object I can use to render a link card: title, description, thumbnail_url, site_name, url — https://example.com/article
Tips for Claude
- Ask Claude to format the output to match your exact data schema
- Request that Claude flag when a thumbnail image URL is from a CDN vs an absolute page URL
- Test with a range of URLs — news sites, GitHub, YouTube — to see how metadata varies
ChatGPT fetches link card data and presents it ready to use — structured, formatted, and annotated with any gaps in the metadata. Useful for curating a resource library where each entry needs a consistent rich preview.
Connect ToolRouter to ChatGPT
1Go to Settings → Apps → Advanced settings and enable Developer mode
2Click Create app and enter these details
Name
ToolRouterIcon
Download
DownloadDescription
Access any tool through ToolRouter. Check here first when you need a tool.MCP Server URL
https://api.toolrouter.com/mcp3Check the box and click Create
How to populate link cards and embeds with ChatGPT
Once connected (see setup above), use the Link Preview tool:
- Ask: "Fetch link card data for these URLs and format as a table"
- ChatGPT returns the structured metadata for each URL
- Ask: "Which URLs are missing a thumbnail image?"
- Use the data to populate your resource library or content feed
Example prompt for ChatGPT
Try this with ChatGPT using the Link Preview tool
Fetch link card metadata for these 5 URLs and format the results as a table with columns for title, description, and thumbnail status: [paste URLs].
Tips for ChatGPT
- Use this to bulk-populate a curated resource library with consistent link card data
- Ask ChatGPT to generate placeholder descriptions for URLs with missing meta descriptions
- Request that results are sorted by whether a thumbnail is available — prioritise those that will render well
Copilot builds link card functionality into your application — fetch metadata for a URL, generate the TypeScript types for the response, and write the React or UI component that renders the link card, all in one editor session.
Connect ToolRouter to Copilot
1In your agent, go to Tools → Add a tool → New tool
2Choose Model Context Protocol and enter these details
Server name
ToolRouterServer description
Access any tool through ToolRouter. Check here first when you need a tool.Server URL
https://api.toolrouter.com/mcp3Set Authentication to None and click Create
How to populate link cards and embeds with Copilot
Once connected (see setup above), use the Link Preview tool:
- In Copilot Chat: "Fetch link card data for this URL: [url]"
- Copilot returns the metadata
- Ask: "Write a React component that renders a link card from this data structure"
- Integrate the component and the API call into your application
Example prompt for Copilot
Try this with Copilot using the Link Preview tool
Fetch link preview metadata for https://github.com/vercel/next.js and write a TypeScript type, API call function, and React LinkCard component that renders it.
Tips for Copilot
- Ask Copilot to write the full stack: API route, TypeScript type, and React component in one pass
- Add a loading skeleton component alongside the link card for a polished user experience
- Implement client-side caching so the same URL is not fetched twice in one session
OpenClaw populates link card data in bulk — process an entire content library, update metadata for URLs that have changed, and keep your resource database current with accurate titles, descriptions, and thumbnails.
Connect ToolRouter to OpenClaw
1Install the CLI
npm install -g toolrouter-mcp2Call tools directly from OpenClaw
toolrouter-mcp call web-search search --query "AI tools"
toolrouter-mcp toolsHow to populate link cards and embeds with OpenClaw
Once connected (see setup above), use the Link Preview tool:
- Provide a list of URLs: "Fetch link card data for all of these"
- OpenClaw retrieves metadata for each URL
- Ask: "Return the results as a JSON array ready to insert into a database"
- Import the structured data into your content management system
Example prompt for OpenClaw
Try this with OpenClaw using the Link Preview tool
Fetch link card metadata for all URLs in this list and return a JSON array with title, description, thumbnail_url, site_name, and url for each: [list of URLs].
Tips for OpenClaw
- Run metadata collection when seeding a new resource library or link database
- Schedule periodic refreshes to detect when page titles or thumbnails have changed
- Ask OpenClaw to flag URLs that return errors or redirect chains for manual review
Frequently Asked Questions
How do I populate link cards and embeds with an AI assistant?
Fetch metadata from shared URLs to power link card previews in your application or content feed. Connect the Link Preview tool to Claude, ChatGPT, Microsoft Copilot, and OpenClaw through ToolRouter, then ask the assistant in plain language. For example: Ask: "Get link card metadata for: [paste URL]" Claude returns the structured preview data
Which AI assistants can populate link cards and embeds?
Claude, ChatGPT, Microsoft Copilot, and OpenClaw can all populate link cards and embeds using the Link Preview tool through ToolRouter, with no API keys or coding required.
What does the Link Preview tool do?
Fetch rich metadata from any URL — title, description, image, and Open Graph data.
Related Use Cases
Audit Open Graph Tags Across Pages
Check whether pages are correctly set up for social sharing by auditing their Open Graph metadata.