web_fetch

Fetch the content of a URL and return it as clean markdown.

web_fetch

Fetches the content of a single URL and returns it as clean markdown. Use to read a specific page (article, doc, blog post) — usually after web_search has surfaced a URL worth examining.

Parameters

Name	Type	Required	Description
url	string (url)	yes	The full http(s) URL to fetch.
max_chars	integer (≥500)	no	Max characters of markdown to return. Default 50000 (~12.5k tokens).

Returns

A formatted markdown payload with the page contents. On failure:

{
  "error": "web_fetch failed: ...",
  "details": {
    "url": "https://example.com/...",
    "http_status": 403,
    "hint": "..."
  }
}

Call from your own custom tool

export async function handle(input, ctx) {
  const page = await ctx.tools.web_fetch({
    url: input.article_url,
    max_chars: 8000,
  });
  return { page };
}

async def handle(input, ctx):
    page = await ctx.tools.web_fetch({
        'url': input['article_url'],
        'max_chars': 8000,
    })
    return { 'page': page }

Notes

• Pages behind paywalls, robot blocks, or that error upstream will return an error envelope.