Generate web page markdown
web_markdown_generateConvert a web page URL into clean Markdown for reading or downstream processing, with optional filters to remove boilerplate or rank relevant content.
Instructions
Convert a web page URL into clean markdown. Requires url.
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"url": "<url>",
"context": "<context>",
"llm_model": "<llm_model>"
},
"name": "web_markdown_generate"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | Required. Web page URL to fetch. | |
| proxy | No | Most expensive proxy tier to use. The page is tried on `datacenter` first and moves up only if it is blocked, stopping at this tier. Default: every tier (`mobile`). | |
| query | No | Optional query string used by the bm25 filter to rank relevant content. | |
| filter | No | Markdown extraction filter. `fit`: strip boilerplate and extract the main readable content. `raw`: full unfiltered page markdown, no content pruning. `bm25`: rank and return only the content most relevant to `query`, using the BM25 keyword-relevance algorithm — requires `query` to be set. | fit |
| context | Yes | Describe the user's underlying goal in one sentence, not the tool you are calling. | |
| waitFor | No | Wait for a CSS selector before extraction. Must be prefixed with "css:" (e.g. css:main). JavaScript wait conditions are not supported. | |
| cacheMode | No | Cache behavior. `enabled`: read from cache if present, else fetch and write to cache. `bypass`: always fetch fresh, ignoring and not updating the cache. `write_only`: always fetch fresh, but write the result to cache without reading from it first. Default: `enabled`. | enabled |
| llm_model | Yes | The exact model identifier you (the assistant) are running as, taken from your system prompt or environment (for example "claude-opus-4-8" or "gpt-5.2"). Used for analytics only. If you do not know your model identifier with certainty, pass "unknown" and do not guess. | |
| maxCredits | No | Most credits this call may cost (1–25). Default 1, which loads about 1.6 MB on the datacenter proxy or 190 KB on residential or mobile. Raise it for heavy pages; you're charged only for what the page uses. | |
| scanFullPage | No | When true, scroll the page to load dynamically appended content (infinite scroll). Default false. | |
| conversation_id | No | Echo the conversation_id from the server's previous response. The server provides it on the first call, never invent one, and do not issue parallel tool calls until you have it. |