Agent Skills

Ready-to-copy integrations for AI agents and LLM frameworks

MCP Server

Connect AI assistants directly to FileShiftr using the Model Context Protocol.

Configuration (Claude Desktop / Cursor / VS Code)

{
  "mcpServers": {
    "fileshiftr": {
      "url": "https://fileshiftr.app:8001/mcp",
      "headers": {
        "Authorization": "Bearer cvt_your_key_here"
      }
    }
  }
}

Works with Claude Desktop, Cursor, Windsurf, and any MCP-compatible client.

Available Tools

list_formats List all supported conversion format pairs (input and output formats).

convert_file Convert a file from one format to another. Accepts base64-encoded input and returns a download URL.

check_conversion_status Poll the status of a conversion job by its ID. Returns status, progress, and download URL when complete.

REST API (Python)

Use the REST API directly from Python scripts or agent tools.

import httpx, time

API_KEY = "cvt_your_key_here"
BASE = "https://fileshiftr.app/api/v1"

# 1. Upload & convert
with open("document.pdf", "rb") as f:
    resp = httpx.post(
        f"{BASE}/convert/",
        headers={"X-API-Key": API_KEY},
        files={"file": ("document.pdf", f)},
        data={"output_format": "md"},
    )
job = resp.json()

# 2. Poll until done
while job["status"] == "processing":
    time.sleep(2)
    job = httpx.get(
        f"{BASE}/conversions/{job['id']}/",
        headers={"X-API-Key": API_KEY},
    ).json()

# 3. Download result
result = httpx.get(
    f"{BASE}/conversions/{job['id']}/download/",
    headers={"X-API-Key": API_KEY},
)
with open("document.md", "wb") as f:
    f.write(result.content)

REST API (TypeScript)

Use the REST API directly from TypeScript with the Fetch API.

const API_KEY = "cvt_your_key_here";
const BASE = "https://fileshiftr.app/api/v1";

// 1. Upload & convert
const form = new FormData();
form.append("file", file);
form.append("output_format", "md");

const { id, status } = await fetch(`${BASE}/convert/`, {
  method: "POST",
  headers: { "X-API-Key": API_KEY },
  body: form,
}).then(r => r.json());

// 2. Poll until done
let job = { id, status };
while (job.status === "processing") {
  await new Promise(r => setTimeout(r, 2000));
  job = await fetch(`${BASE}/conversions/${job.id}/`, {
    headers: { "X-API-Key": API_KEY },
  }).then(r => r.json());
}

// 3. Download result
const blob = await fetch(`${BASE}/conversions/${job.id}/download/`, {
  headers: { "X-API-Key": API_KEY },
}).then(r => r.blob());

OpenAI Function Calling

Define FileShiftr as a tool for GPT-based agents.

Tool Definition

{
  "type": "function",
  "function": {
    "name": "convert_file",
    "description": "Convert a file from one format to another using FileShiftr. Supports 39+ format pairs including PDF, DOCX, images, spreadsheets, and AI-powered conversions.",
    "parameters": {
      "type": "object",
      "properties": {
        "file_path": {
          "type": "string",
          "description": "Path to the input file"
        },
        "output_format": {
          "type": "string",
          "description": "Target format (e.g. md, pdf, jpg, png, docx, html, yaml, json, csv, xlsx)"
        }
      },
      "required": ["file_path", "output_format"]
    }
  }
}

Implementation (handles the tool call)

import httpx, time

API_KEY = "cvt_your_key_here"
BASE = "https://fileshiftr.app/api/v1"

def handle_convert_file(file_path: str, output_format: str) -> str:
    """Execute the convert_file tool call via the FileShiftr REST API."""
    with open(file_path, "rb") as f:
        resp = httpx.post(
            f"{BASE}/convert/",
            headers={"X-API-Key": API_KEY},
            files={"file": (file_path.split("/")[-1], f)},
            data={"output_format": output_format},
        )
    job = resp.json()

    while job["status"] == "processing":
        time.sleep(2)
        job = httpx.get(
            f"{BASE}/conversions/{job['id']}/",
            headers={"X-API-Key": API_KEY},
        ).json()

    if job["status"] != "done":
        return f"Conversion failed: {job.get('error_message', 'unknown error')}"

    result = httpx.get(
        f"{BASE}/conversions/{job['id']}/download/",
        headers={"X-API-Key": API_KEY},
    )
    out_path = file_path.rsplit(".", 1)[0] + f".{output_format}"
    with open(out_path, "wb") as f:
        f.write(result.content)
    return f"Converted to {out_path}"

Claude Tool Use (Anthropic API)

Define FileShiftr as a tool for Claude-based agents.

Tool Definition

{
  "name": "convert_file",
  "description": "Convert a file from one format to another using FileShiftr. Supports 39+ format pairs including PDF, DOCX, images, spreadsheets, and AI-powered conversions.",
  "input_schema": {
    "type": "object",
    "properties": {
      "file_path": {
        "type": "string",
        "description": "Path to the input file"
      },
      "output_format": {
        "type": "string",
        "description": "Target format (e.g. md, pdf, jpg, png, docx, html, yaml, json, csv, xlsx)"
      }
    },
    "required": ["file_path", "output_format"]
  }
}

LangChain

Use FileShiftr as a LangChain tool in your agent chains.

from langchain.tools import tool
import httpx, time

API_KEY = "cvt_your_key_here"
BASE = "https://fileshiftr.app/api/v1"

@tool
def convert_file(file_path: str, output_format: str) -> str:
    """Convert a file to another format using FileShiftr.
    Supports: PDF, DOCX, PPTX, XLSX, images, JSON, YAML, CSV, XML, HTML, Markdown.
    """
    with open(file_path, "rb") as f:
        resp = httpx.post(
            f"{BASE}/convert/",
            headers={"X-API-Key": API_KEY},
            files={"file": (file_path.split("/")[-1], f)},
            data={"output_format": output_format},
        )
    job = resp.json()

    while job["status"] == "processing":
        time.sleep(2)
        job = httpx.get(
            f"{BASE}/conversions/{job['id']}/",
            headers={"X-API-Key": API_KEY},
        ).json()

    if job["status"] != "done":
        return f"Conversion failed: {job.get('error_message', 'unknown error')}"

    result = httpx.get(
        f"{BASE}/conversions/{job['id']}/download/",
        headers={"X-API-Key": API_KEY},
    )
    out_path = file_path.rsplit(".", 1)[0] + f".{output_format}"
    with open(out_path, "wb") as f:
        f.write(result.content)
    return f"Converted to {out_path}"

Get your free API key

Free tier includes 5 conversions/day. Upgrade to Premium for unlimited access.