inspect_url
Fetches a URL and returns clean markdown or text plus metadata, without browser clutter.
DESCRIPTION
Use this when an agent needs external documentation or article content in a compact format that is actually readable in context.
WHEN TO USE
GOOD FIT
- Pulling content from documentation pages.
- Extracting a clean article or API reference for an agent to read.
- Comparing external docs without opening a browser session.
AVOID IT WHEN
- You need TLS and security header analysis; use audit_headers.
- You need interactive browser behavior.
- You need local repo analysis instead of external content.
PARAMETERS
| Name | Type | Required | Description |
|---|---|---|---|
| url | string | Yes | URL to inspect. |
| format | "markdown" | "text" | No | Response format. "markdown" is the default. |
EXAMPLE
tools-call.json
{ "jsonrpc": "2.0", "id": 8, "method": "tools/call", "params": { "name": "inspect_url", "arguments": { "url": "https://example.com/docs", "format": "markdown" } }}OUTPUT
response.json
{ "jsonrpc": "2.0", "id": 8, "result": { "content": [ { "type": "text", "text": "# Example Docs\n\nUpdated: 2026-04-05\n\n## Authentication\nSend Bearer auth on every request..." } ] }}TOKEN EFFICIENCY
Because it strips browser chrome and noise, this tool usually gives a tighter context payload than pasting raw HTML or full rendered pages into an agent.