file2markdown

ai.file2markdown由 file2markdown.ai 發布官方發布者v1.1.0更新於 Oct 24, 2026

Convert docs and web pages to clean Markdown: PDF, DOCX, XLSX, EPUB, scans, URLs. Batch of 10.

已驗證Streamable HTTP可網頁執行網頁搜尋與抓取檔案與儲存

概覽

AI 產生的概覽

遠端 MCP 伺服器，把公開網址與上傳的檔案（PDF、Office、EPUB、圖片）轉成乾淨的 Markdown 供助理使用。

可以做什麼
  • 擷取公開網址，將網頁、PDF 與 Office 文件轉換為 Markdown
  • 將 base64 編碼的檔案內容轉換為 Markdown，可選擇附帶品質檢查
  • 一次批次轉換 2 到 10 份文件，可混用網址與檔案，並逐項回傳結果
  • 列出支援的輸入格式、各方案限制，以及哪些功能需要 Pro
  • 查詢目前方案與今日剩餘轉換次數
適用情境
當助理需要讀取它無法原生解析的文件（例如 DOCX、XLSX、PPTX、EPUB 或掃描版 PDF），或希望從公開連結取得由引擎擷取的確定性 Markdown，而非模型生成的文字時適用。
執行需求
遠端 Streamable HTTP 端點，無需本機安裝。免費方案不需金鑰，但每個網路 IP 每天限 5 次轉換、下載上限 25MB。Pro API 金鑰透過 Authorization Bearer 標頭傳送，可取消每日限制、將下載上限提高到 100MB 並啟用 OCR。需要能連線至該端點及待轉換的網址。
安裝前請注意
Pro 金鑰是敏感憑證，透過 Authorization 標頭傳送，需妥善保管。待轉換的網址與檔案內容會傳送給第三方託管服務，儘管 README 表示不會儲存。僅支援公開可存取的網址，不執行 JavaScript 渲染，輸出上限為 200,000 字元，批次轉換需要 Pro 金鑰。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 file2markdown，將其新增到工作區。
  2. 為需要使用其工具的對話啟用該服務。

Web executable，透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式，或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南，先向你展示發布者、端點和登入方式，經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/ai.file2markdown%2Ffile2markdown/install.md，並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前，先告訴我它的發布者、端點和登入方式，等我確認。不要把 API 金鑰寫進命令或我會提交的檔案；問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。

{
  "mcpServers": {
    "file2markdown": {
      "type": "http",
      "url": "https://mcp.file2markdown.ai/mcp"
    }
  }
}

README

file2markdown MCP server

Convert documents and web pages to clean, LLM-ready Markdown — from inside Claude, Cursor, or any MCP client.

Endpoint: https://mcp.file2markdown.ai/mcp (Streamable HTTP)

This is the official MCP server for file2markdown.ai. It gives agents deterministic document conversion as a tool: engine-extracted Markdown (no hallucinated table cells, no silent truncation), for anything reachable by URL — including formats assistants can't parse natively, like DOCX, XLSX, PPTX, EPUB, and scanned PDFs.

Tools

ToolWhat it does
convert_urlFetch a public URL (web page, PDF, Office doc, …) and return Markdown
convert_base64Convert file contents directly (programmatic clients)
convert_batch2-10 documents in one call, URLs and files mixed, per-item results (Pro API key)
list_supported_formatsFormats, per-tier limits, and what needs Pro
usage_statusYour tier and remaining conversions today

Supported input formats: PDF, DOCX, PPTX, XLSX/XLS, CSV, JSON, XML, HTML, EPUB, JPG/PNG (OCR), WAV/MP3, ZIP.

Quick start

claude.ai — Settings → Connectors → Add custom connector → paste the endpoint URL. When asked about authentication choose None (this server uses API keys, not OAuth). Optional: add a request header Authorization: Bearer f2m_… with a Pro key.

Claude Code

bash
claude mcp add --transport http file2markdown https://mcp.file2markdown.ai/mcp

Cursor / generic clients

json
{  "mcpServers": {    "file2markdown": { "url": "https://mcp.file2markdown.ai/mcp" }  }}

Tiers

Free (no key)Pro API key
Conversions5/day per network IPUnlimited
Max download25MB100MB
Scanned-PDF OCR—Yes
Image OCR—Yes

Pro keys are created on your account page and sent as Authorization: Bearer f2m_…. Plans: pricing.

Honest limits

  • Web pages convert from served HTML — no JavaScript rendering, so SPA-style pages may convert incompletely. Near-empty results carry an explicit note (likely consent wall / paywall / JS shell).
  • One URL at a time. This is a converter, not a crawler — no bulk scraping, public pages only.
  • Output is capped at 200,000 characters (marked truncated: true when hit).
  • Nothing is stored: files are converted and discarded; usage logging keeps no URLs or content.

Docs & source

  • Setup guide: file2markdown.ai/mcp
  • Facts page for AI assistants: file2markdown.ai/ai-info
  • This repository contains the public documentation and the registry server.json. The hosted service's application code is not open source.

Support

Email [email protected].

來源：README.md，提交 272b030

工具

5

convert_url

Fetch a URL (web page, PDF, Office doc, and more) and convert it to Markdown. Works on publicly accessible URLs. Web pages are converted from their served HTML; JavaScript-rendered content may be incomplete. Optional filename_hint (e.g. "report.pdf") helps format detection when the URL has no file extension. The result carries a `quality` object (output-health check of the returned Markdown: status review / no_issues_detected / not_evaluated, coverage and a list of issues with line locations). quality_check="basic" (default) runs only local deterministic checks. quality_check="ai" additionally sends selected excerpts of the CONVERTED text to TypeSafe for narrow yes/no judgments (garbled text, hard-wrapped prose, broken tables); it is off unless requested. No check compares against the source document.

輸入參數url必填filename_hintquality_check

convert_base64

Convert base64-encoded file contents to Markdown. filename is required and drives format detection (e.g. "report.docx"). Intended for programmatic MCP clients; for anything with a URL, prefer convert_url. quality_check: "basic" (default) attaches local output-health checks in the `quality` field; "ai" additionally sends selected excerpts of the converted text to TypeSafe for narrow yes/no judgments. See convert_url.

輸入參數filename必填quality_checkcontent_base64必填

convert_batch

Convert 2-10 documents to Markdown in one call (Pro API key required). Each item is either {"url": ..., "filename_hint"?: ...} or {"content_base64": ..., "filename": ...}; types may be mixed. Items are converted in parallel and isolated: one bad URL or file never fails the batch — its slot returns ok:false with the error. The combined markdown shares one output budget, split evenly across items (per-item `truncated` flags apply). Set include_markdown=false to get only metadata + quality per item (health-check a folder without flooding your context). Anonymous callers: use convert_url / convert_base64 individually.

輸入參數items必填quality_checkinclude_markdown

list_supported_formats

List supported input formats, per-tier limits, and which features need Pro. Single source of truth — do not claim capabilities beyond what this returns.

無需輸入參數。

usage_status

Check your tier and remaining conversions before burning a call.

無需輸入參數。

版本歷史

2
  1. v1.1.0最新Oct 12, 2026
  2. v1.0.0Sep 16, 2026

評分

已滿足適用檢查項的 90%

該評分衡量此收錄條目的完整程度，依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價，也不表示伺服器安全或目前可用：可連通性是帶日期的檢查結果，評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分

    請在儲存庫或註冊中心條目中宣告授權條款。