---
title: "CLI reference"
description: "Every webfold subcommand, with its arguments."
---

This page mirrors the usage text compiled into the binary. For the version you have installed, run `webfold --help`. MCP server mode is the separate `webfold-mcp` binary.

## Core extraction

| Command | Arguments | Does |
| --- | --- | --- |
| `scrape` | `<url\|file> [base] [--chunk n] [--json] [--action-budget n] [--timeout n]` | Acquire and convert to Markdown. |
| `scrape --render` | `<url> [--recovery-json] [--wait-for text] [--wait-selector css] [--browser endpoint] [--screenshot] [--scroll-document] [--expand-disclosures css] [--document-selector css] [--include-link-labels] [--include-aria-hidden] [--include-form-text]` | Extract through a Chromium you run. See [Rendered pages](/docs/guides/rendering). |
| `md` | `<file> [base] [--chunk n] [--json]` | Raw HTML to Markdown, no fetch. |
| `crawl` | `<dir\|url> [--min n] [--max n] [--chunk n]` | Per-page content with chrome folded once. |
| `fetch` | `<url> [--dump text\|html\|links] [--quiet] [--timeout N]` | Fetch only; print text, HTML or links. |
| `convert` | `<document-file>` | Convert a document to Markdown. |
| `chunk` | `<html\|md\|txt\|pdf\|url> [--size n] [--jsonl]` | Retrieval-ready chunks. |
| `pdf` | `<file.pdf> [--chunk n]` | PDF to Markdown. |
| `webmcp` | `<url\|file> [--tool name --input json]` | List declarative WebMCP form tools on a page, or validate a call into a fill plan. |

## Search and indexing

| Command | Arguments | Does |
| --- | --- | --- |
| `search` | `<query> [--count n] [--backend ddg\|bing\|ecosia\|auto] [--consensus [--profile web\|tech\|academic\|all \| --engines a,b,c]] [--corpus dir]` | Search the web, several engines, or a saved corpus. |
| `serp` | `<engine> <html-file> [base-url]` | Normalise a saved results page. |
| `index crawl` | `<seed-url> [--max n] [--depth n] [--out file] [--http-cache dir]` | Crawl and serialise an index. |
| `index search` | `<seed-url> <query> [--max n] [--depth n] [--top n]` | Crawl, then rank pages for a query. |
| `index map` | `<seed-url> [--max n] [--no-crawl-fallback]` | Discover same-origin URLs. |
| `index resolve` | `<guessed-url> [--max n]` | Resolve a guessed URL to the canonical one. |
| `folder-search` | `<dir> <query> [-n k] [--regex] [--semantic] [--no-cache] [--stats]` | Search a local folder with a memory-mapped index and phrase ranking. |
| `folder-search serve` | `<dir> [--socket path] [--refresh-ms n]` | Keep the index mapped in one process and answer one JSON line per query, on stdin or a Unix socket. |
| `dataset-url`, `dataset-extract` | `<query>` / `<file.html> [--url u]` | Dataset Search URL; schema.org Dataset records. |

## Notes, sinks, jobs

| Command | Arguments | Does |
| --- | --- | --- |
| `note` | `<url\|file> --vault <dir> [--folder name] [--title text]` | Save a page as a note in a vault. |
| `sink` | `<jsonl\|parquet\|iceberg> [options] < records.ndjson` | Write NDJSON records to a sink. |
| `job` | `run\|watch <url> [--checks seo,tells,pii,brand] [--interval secs]` | Run analysis checks once or on an interval. |
| `workflow` | `<workflow.json>` | Run a workflow file. |

## Browser and network

| Command | Arguments | Does |
| --- | --- | --- |
| `headless run` | `<url> [--stealth] [--proxy host:port]` | One-shot fetch through headless mode. |
| `headless serve` | `[port] [--stealth] [--profile DIR]` | CDP server (default port 9222). |
| `headless capabilities` |  | JSON: CDP methods implemented, no-op, unsupported. |
| `proxy` | `[port]` | Intercepting forward proxy (default 8080). |

## Social

| Command | Arguments |
| --- | --- |
| `social platforms` |  |
| `social read` | `<url> [--jsonl]` |
| `social search` | `<platform> <query> [--jsonl]` |
| `social crawl` | `<url> [--max n] [--jsonl]` |
| `social trajectory` | `<url>...` |

## WARC and batch

| Command | Arguments |
| --- | --- |
| `warc` | `<file.warc\|.warc.gz> [--threads n] [--max n]` |
| `dump` | `<warc-file> <out-file> [--max n]` |
| `replay` | `<file.warc\|.warc.gz> [--out replay.html] [--max n] [--full-html]` |
| `batch` | `<dir-of-html-files>` |
| `capture` | `<url> -o <out.warc\|.warc.gz> [--backend native] [--tlp RED\|AMBER\|AMBER+STRICT\|GREEN] [--retention-days N] [--no-gzip]` |

## Reachability and diagnostics

| Command | Arguments | Does |
| --- | --- | --- |
| `verify` | `<url>... \| --file urls.txt [--parallel n] \| --jsonl file \| --diff new old` | Probe URLs through the native backend. JSON Lines out: status, content type, latency, redirect count, error. |
| `doctor` | `[--json]` | Build profile, pinned dependency revisions, git HEAD, environment check. No I/O. |
| `--version` |  | Print the version. |

The repository also ships `osint`, `view`, `totp` and `login` commands behind default features. Run `webfold --help` on your build for the full list.