Convert WebVTT to Markdown

Upload a .vtt subtitle file and get clean, flowing text back instantly — including speaker labels, with timecodes stripped out. No installation, no account required.

Why convert VTT to Markdown?

A .vtt file is mostly timecodes and markup tags that mean nothing to a language model but still cost tokens. slim-down turns the separate subtitle lines into flowing paragraphs with speaker labels, so models like ChatGPT, Claude, and Gemini can read the content as a coherent transcript.

How does it work?

Upload your .vtt file or drag it into the converter.
slim-down strips the timecodes, recognizes speakers from WebVTT voice tags, and groups the text into paragraphs based on pauses and speaker changes.
Download or copy the resulting Markdown, ready to paste straight into a prompt.

Open Converter

Frequently Asked Questions

Are the timecodes from my VTT file preserved?

No, by design: timecodes add nothing for an AI model and only cost tokens. The text is turned into flowing paragraphs, starting a new one after a pause of more than three seconds or a speaker change.

Does slim-down recognize who's speaking?

Yes. If your .vtt file uses WebVTT voice tags with a speaker name, slim-down converts them into a **Speaker Name:** label before the matching text, and merges repeated auto-caption lines.

I have an .srt file, can I convert that too?

Yes, SRT files (.srt) are converted the same way into flowing Markdown text, though they typically don't carry speaker labels.

Is converting VTT to Markdown free?

Yes. Without an account you get up to 15 free conversions per day. A free account gives you unlimited conversions and lets you save and sync documents across devices.

🔑 REST API Access & Keys

Not logged in yet

Securely sync documents across all your devices.

⚡ REST API Endpoints

POST /api/convert

Tekst, CSV, JSON, HTML, YAML

Converts input to Markdown or another target format. Optionally add save: true to automatically save the result to your account.

curl -X POST https://slim-down.nl/api/convert \
  -H "Authorization: Bearer <JOUW_API_SLEUTEL>" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "Product,Prijs\nLaptop,999",
    "source_format": "csv",
    "target_format": "markdown"
  }'

POST /api/convert-file

.pdf, .docx, .doc, .pptx, .ppt, .xlsx, .xls, .ipynb, .srt, .vtt

Upload PDF, Word, PowerPoint, Excel, Jupyter Notebook, or subtitle (.srt/.vtt) files via multipart form-data for direct conversion to Markdown or CSV.

curl -X POST https://slim-down.nl/api/convert-file \
  -H "X-API-Key: <JOUW_API_SLEUTEL>" \
  -F "file=@document.docx"

MCP /mcp

Claude Desktop, Claude Code, agents

Use the same API key to connect slim-down as an MCP tool — the agent calls convert and convert_file directly, with no REST boilerplate. For Claude Desktop and more detail, see the MCP Server page.

claude mcp add --transport http slim-down https://slim-down.nl/mcp \
  --header "Authorization: Bearer <JOUW_API_SLEUTEL>"

GET /api/documents

Cloud Sync

Retrieves all your saved documents in JSON format for sync across apps and devices.

curl -X GET https://slim-down.nl/api/documents \
  -H "Authorization: Bearer <JOUW_API_SLEUTEL>"

GET /api/auth/me

User Info & Token Validation

Checks if your API key is valid and returns user profile information.

curl -X GET https://slim-down.nl/api/auth/me \
  -H "Authorization: Bearer <JOUW_API_SLEUTEL>"

🤖 AI Integrations & Token Reduction Examples

Claude • Gemini • Copilot

Convert documents or raw data via slim-down to compact Markdown before sending them as context to an LLM to save tokens and costs.

⚡ Claude Tool / Skill (Function Calling) (Anthropic Tool Use schema & Python handler)

Give Claude the skill/tool to automatically call the slim-down API when it needs to simplify a document, table, or JSON/HTML:

Anthropic Console ↗

import requests
from anthropic import Anthropic

SLIM_DOWN_KEY = "JOUW_SLIM_DOWN_API_SLEUTEL"
SLIM_DOWN_URL = "https://slim-down.nl/api/convert"

# 1. Definieer de slim-down Skill / Tool voor Claude
tools = [
    {
        "name": "slim_down_convert",
        "description": "Converteert HTML, JSON, YAML, CSV, Word of platte tekst naar compacte Markdown om tokens te besparen.",
        "input_schema": {
            "type": "object",
            "properties": {
                "content": {
                    "type": "string",
                    "description": "De ruwe tekst, html, json, yaml of csv content"
                },
                "target_format": {
                    "type": "string",
                    "enum": ["markdown", "json", "yaml", "csv", "excel", "text"],
                    "default": "markdown"
                }
            },
            "required": ["content"]
        }
    }
]

# 2. Tool handler functie
def run_slim_down_tool(content: str, target_format: str = "markdown") -> str:
    resp = requests.post(
        SLIM_DOWN_URL,
        headers={"Authorization": f"Bearer {SLIM_DOWN_KEY}"},
        json={"input": content, "target_format": target_format}
    )
    return resp.json().get("result", "")

# 3. Claude aanroepen met de tool
client = Anthropic()
response = client.messages.create(
    model="claude-3-7-sonnet-20250219",
    max_tokens=1024,
    tools=tools,
    messages=[
        {
            "role": "user",
            "content": "Kun je deze HTML tabel versimpelen naar Markdown?\n<table><tr><th>Naam</th><th>Score</th></tr><tr><td>Alex</td><td>95</td></tr></table>"
        }
    ]
)
print("Tool calls:", response.content)

💬 System Prompts & AI Instructions (Claude Projects, Custom GPTs & Cursor rules)

Use this prompt in your Claude Project Custom Instructions, System Prompt, or .cursorrules:

Claude Projects ↗ • Custom GPTs ↗

### SYSTEM PROMPT / AI INSTRUCTION:
You have access to the slim-down conversion API at https://slim-down.nl.
When the user asks to analyze large HTML pages, Word files (.docx), Excel sheets (.xlsx), JSON data, or CSV:
1. Use the slim-down API endpoint (POST /api/convert or POST /api/convert-file)
   with header "Authorization: Bearer <YOUR_SLIM_DOWN_KEY>".
2. Always convert the input to "markdown" first to minimize token consumption.
3. Answer the user's question solely based on the resulting clean Markdown.

🟣 Anthropic Claude (Direct SDK) (Python Anthropic SDK + Document Context)

First upload your Word/Excel/JSON file to slim-down and send the clean Markdown context directly to Claude 3.5 / 3.7:

Anthropic Docs ↗

import requests
from anthropic import Anthropic

SLIM_DOWN_KEY = "JOUW_SLIM_DOWN_API_SLEUTEL"
CLAUDE_KEY = "sk-ant-..."

# 1. Converteer Word/Excel naar compacte Markdown via slim-down
with open("jaarverslag.docx", "rb") as f:
    res = requests.post(
        "https://slim-down.nl/api/convert-file",
        headers={"X-API-Key": SLIM_DOWN_KEY},
        files={"file": f}
    )
clean_markdown = res.json()["result"]

# 2. Stuur de geoptimaliseerde context naar Claude
client = Anthropic(api_key=CLAUDE_KEY)
message = client.messages.create(
    model="claude-3-7-sonnet-20250219",
    max_tokens=1000,
    messages=[
        {
            "role": "user",
            "content": f"Analyseer dit document en geef de 3 belangrijkste KPI's:\n\n{clean_markdown}"
        }
    ]
)
print(message.content[0].text)

🔵 Google Gemini (Python google-genai / google-generativeai SDK)

Optimize spreadsheets or raw tables into Markdown tables for Gemini 2.0 / 1.5 Flash:

Google AI Studio ↗

import requests
import google.generativeai as genai

SLIM_DOWN_KEY = "JOUW_SLIM_DOWN_API_SLEUTEL"
genai.configure(api_key="AIzaSy...")

# 1. Converteer complexe spreadsheet of CSV data naar Markdown tabel
raw_csv = (
    "Afdeling,Q1,Q2,Budget\n"
    "Sales,120000,145000,250000\n"
    "Marketing,45000,60000,100000"
)
conv = requests.post(
    "https://slim-down.nl/api/convert",
    headers={"Authorization": f"Bearer {SLIM_DOWN_KEY}"},
    json={"input": raw_csv, "source_format": "csv", "target_format": "markdown"}
).json()

# 2. Vraag Gemini om samenvatting met minimale tokens
model = genai.GenerativeModel("gemini-2.0-flash")
response = model.generate_content(
    f"Hier zijn de kwartaalcijfers in Markdown:\n\n{conv['result']}\n\nWelke afdeling benadert zijn budget?"
)
print(response.text)

🤖 GitHub Copilot, Cursor & ChatGPT (Prompt Engineering & Context Injection)

Copy the converted Markdown directly into your IDE or Copilot Chat window with the following structured prompt template:

OpenAI Platform ↗ • Copilot ↗

You are an AI assistant. Use the specification below (converted via slim-down)
as the source of truth for the implementation.

### SOURCE DOCUMENT (Markdown):
{{PASTE CONVERTED SLIM-DOWN OUTPUT HERE}}

### INSTRUCTIONS:
Write the required Go / TypeScript dataclasses and validation rules based on the tables above.

🐍 🟨 Python & JavaScript Integration Examples

Python (requests)

import requests

headers = {
    "Authorization": "Bearer JOUW_API_SLEUTEL",
    "Content-Type": "application/json"
}
payload = {
    "input": "# Titel\nInhoud converteren",
    "target_format": "html"
}
res = requests.post("https://slim-down.nl/api/convert", json=payload, headers=headers)
print(res.json()["result"])

JavaScript / Node.js (fetch)

const res = await fetch("https://slim-down.nl/api/convert", {
  method: "POST",
  headers: {
    "Authorization": "Bearer JOUW_API_SLEUTEL",
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    input: "Product,Prijs\nBoek,15.00",
    target_format: "markdown"
  })
});
const data = await res.json();
console.log(data.result);