File to Markdown — Skill\r

\r

Overview\r

\r Convert files into clean, structured, AI-ready Markdown using the markdown.new API powered by Cloudflare Workers AI toMarkdown().\r \r Supports 20+ formats including documents, spreadsheets, images, and structured data.\r \r No authentication required (500 requests/day per IP).\r \r ---\r \r

When to Use This Skill\r

\r Use this skill whenever you need to:\r \r

• Extract text from files for LLM processing\r
• Convert PDFs or Office files into Markdown\r
• Normalize data into structured text\r
• Process uploaded user files\r
• Scrape webpage content into Markdown\r
• Convert images into AI-generated descriptions + content\r \r Common AI workflows:\r \r
• RAG ingestion pipelines\r
• Knowledge base creation\r
• Document summarization\r
• Dataset extraction\r
• Spreadsheet analysis\r
• OCR-like extraction from images\r \r ---\r \r

Supported Formats\r

\r

Documents\r

\r

• .pdf\r
• .docx\r
• .odt\r \r

Spreadsheets\r

\r

• .xlsx\r
• .xls\r
• .xlsm\r
• .xlsb\r
• .et\r
• .ods\r
• .numbers\r \r

Images\r

\r

• .jpg\r
• .jpeg\r
• .png\r
• .webp\r
• .svg\r \r

Text & Structured Data\r

\r

• .txt\r
• .md\r
• .csv\r
• .json\r
• .xml\r
• .html\r
• .htm\r \r Notes:\r \r
• Image conversion uses AI object detection + summarization.\r
• HTML URL conversion uses a web page pipeline.\r
• Uploaded HTML uses Workers AI conversion.\r \r ---\r \r

API Base URL\r

\r

https://markdown.new\r
```\r
\r
---\r
\r
## Endpoints\r
\r
### 1️⃣ Convert Remote File (Simple GET)\r
\r
Returns plain Markdown text.\r
\r
```\r
GET /:file-url\r
```\r
\r
Example:\r
\r
```bash\r
curl -s "https://markdown.new/https://example.com/report.pdf"\r
```\r
\r
---\r
\r
### 2️⃣ Convert Remote File (JSON Response)\r
\r
Returns metadata + Markdown.\r
\r
```\r
GET /:file-url?format=json\r
```\r
\r
Example:\r
\r
```bash\r
curl -s "https://markdown.new/https://example.com/report.pdf?format=json"\r
```\r
\r
---\r
\r
### 3️⃣ Convert Remote File via POST\r
\r
Use when you want structured JSON response.\r
\r
```\r
POST /\r
Content-Type: application/json\r
```\r
\r
Body:\r
\r
```json\r
{\r
  "url": "https://example.com/report.pdf"\r
}\r
```\r
\r
Example:\r
\r
```bash\r
curl -s https://markdown.new/ \\r
  -H "Content-Type: application/json" \\r
  -d '{"url": "https://example.com/report.pdf"}'\r
```\r
\r
---\r
\r
### 4️⃣ Upload Local File\r
\r
Use when file is not publicly accessible.\r
\r
```\r
POST /convert\r
multipart/form-data\r
```\r
\r
Example:\r
\r
```bash\r
curl -s https://markdown.new/convert \\r
  -F "[email protected]"\r
```\r
\r
---\r
\r
## Response Formats\r
\r
### URL Conversion Response\r
\r
```json\r
{\r
  "success": true,\r
  "url": "https://example.com/report.pdf",\r
  "title": "Quarterly Report",\r
  "content": "# Quarterly Report\
\
...",\r
  "method": "Workers AI (file)",\r
  "duration_ms": 1200,\r
  "tokens": 850\r
}\r
```\r
\r
---\r
\r
### Upload Conversion Response\r
\r
```json\r
{\r
  "success": true,\r
  "data": {\r
    "title": "Q4 Report",\r
    "content": "# Q4 Report\
\
...",\r
    "filename": "report.xlsx",\r
    "file_type": ".xlsx",\r
    "tokens": 1250,\r
    "processing_time_ms": 320\r
  }\r
}\r
```\r
\r
---\r
\r
## Best Practices for AI Agents\r
\r
### Prefer GET for Simple Workflows\r
\r
Use:\r
\r
```\r
GET /:url\r
```\r
\r
When:\r
\r
* You only need Markdown text\r
* Speed is important\r
* No metadata required\r
\r
---\r
\r
### Prefer POST for Structured Pipelines\r
\r
Use POST when:\r
\r
* Metadata is needed\r
* Token counts are required\r
* Monitoring or logging is implemented\r
* Building automation workflows\r
\r
---\r
\r
### File Upload Strategy\r
\r
Use `/convert` only if:\r
\r
* File is local\r
* File is private\r
* File requires authentication to access\r
\r
Otherwise always prefer URL conversion.\r
\r
---\r
\r
## Error Handling Strategy\r
\r
Agents should:\r
\r
1. Check `"success": true`\r
2. Retry once if network failure\r
3. Validate content length > 0\r
4. Fallback to alternate extraction if needed\r
\r
---\r
\r
## Rate Limits\r
\r
* 500 requests/day per IP without API key\r
* No signup required\r
\r
Agents should:\r
\r
* Cache results when possible\r
* Avoid duplicate conversions\r
\r
---\r
\r
## Integration Examples\r
\r
### JavaScript (Node.js)\r
\r
```js\r
const res = await fetch("https://markdown.new/", {\r
  method: "POST",\r
  headers: { "Content-Type": "application/json" },\r
  body: JSON.stringify({\r
    url: "https://example.com/file.pdf"\r
  })\r
});\r
\r
const data = await res.json();\r
console.log(data.content);\r
```\r
\r
---\r
\r
### Python\r
\r
```python\r
import requests\r
\r
res = requests.post(\r
    "https://markdown.new/",\r
    json={"url": "https://example.com/file.pdf"}\r
)\r
\r
data = res.json()\r
print(data["content"])\r
```\r
\r
---\r
\r
## Agent Decision Tree\r
\r
If user provides:\r
\r
| Input Type      | Action                 |\r
| --------------- | ---------------------- |\r
| Public file URL | Use GET or POST        |\r
| Local file      | Use POST /convert      |\r
| Image           | Convert then summarize |\r
| Spreadsheet     | Convert then analyze   |\r
| Webpage         | Convert URL HTML       |\r
\r
---\r
\r
## Output Expectations\r
\r
The Markdown should be:\r
\r
* Clean\r
* Structured\r
* AI-friendly\r
* Minimal noise\r
* Ready for LLM ingestion\r
\r
---\r
\r
## Limitations\r
\r
* Complex PDF layouts may lose formatting\r
* Large spreadsheets may be truncated\r
* Images rely on AI interpretation accuracy\r
* Token limits may apply\r
\r
---\r
\r
## Summary\r
\r
This skill provides a **universal file-to-Markdown conversion layer** for AI systems with:\r
\r
* No authentication\r
* Simple HTTP interface\r
* Multi-format support\r
* Structured output\r
* Fast processing\r
\r
Ideal for document ingestion, RAG pipelines, and automation agents.\r
\r
---\r