AgenticSchema
Converts webpages into typed JSON, MCP/WebMCP tools, and epistemic context for AI agents.
Tested on Linux
Use it
Linux x86_64
You need first
- Needs the network
install @agenticschema/server globally via npm
npm install -g @agenticschema/server
Where the README and reality differ
- README: npx @agenticschema/server --live starts persistent Live Search MCP server
What worked: @agenticschema/server CLI rejects --live (unknown option); stdio mode requires one or more URLs - README: npx @agenticschema/server --json --fallback / --context --early-abort / --resolution L0 examples
What worked: server package only supports --max-tools, --no-actions, --allow-host, --http, --port, --quiet
Check that it works
agenticschema --helpThe output should contain “agenticschema: serve a page's Schema.org markup as an MCP server”.
Add it to your agent
claude mcp add agenticschema -- agenticschema https://en.wikipedia.org/wiki/BackpackJSON for Claude Desktop, Cursor and other clients
{
"mcpServers": {
"agenticschema": {
"command": "agenticschema",
"args": [
"https://en.wikipedia.org/wiki/Backpack"
]
}
}
}Add agenticschema <url> as an MCP server in your MCP client, giving it the webpage URL you want to work with.
What it does
- Turn webpages into typed JSONTested
How to use it
- Add the server to your MCP client with
agenticschema <url>, using the page you want. - Call the
get_articletool from your client to get the page content as structured data.
- Add the server to your MCP client with
- Expose websites as callable WebMCP toolsNot tested
- Generate epistemic context for AI searchNot tested
How far pickfit got
Linux x86_64
- Installs
- Connects
- 1 of 3 features work
- Needs the network
Good for
- scrape a page to structured JSON
- serve live search evidence via MCP
- print hybrid LLM context for a URL
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
An MCP client called one of its tools
- Replayed from scratch
- Needed the network
- Checked on 31 Oct 2026
- Version 0.0.0
An MCP client handshake over stdio listed 3 tools, and a call to get_article returned a result.