Enables retrieving transcripts and parsing content from YouTube videos for use by AI agents.
MCP (Model Context Protocol) server for the ParseJet API. Lets AI agents like Claude Code and Cursor parse files, URLs, and YouTube transcripts.
Tool | Description |
| Parse any URL (web page, YouTube, PDF link, etc.) |
| Parse a local file (PDF, DOCX, image, etc.) |
| Get a YouTube video transcript |
Related MCP server: md-converter-mcp
npm install -g @parsejet/mcp-serverOr run directly with npx:
npx @parsejet/mcp-serverSet the PARSEJET_API_KEY environment variable to authenticate. Without it, the server works anonymously with a limit of 3 requests per day.
Get your API key at parsejet.com/dashboard.
Add to your project's .claude/settings.json (or global ~/.claude/settings.json):
{
"mcpServers": {
"parsejet": {
"command": "npx",
"args": ["-y", "@parsejet/mcp-server"],
"env": {
"PARSEJET_API_KEY": "your-api-key"
}
}
}
}Add to your Cursor MCP config (Settings > MCP Servers):
{
"mcpServers": {
"parsejet": {
"command": "npx",
"args": ["-y", "@parsejet/mcp-server"],
"env": {
"PARSEJET_API_KEY": "your-api-key"
}
}
}
}If installed globally, you can use parsejet-mcp instead of npx:
{
"mcpServers": {
"parsejet": {
"command": "parsejet-mcp",
"env": {
"PARSEJET_API_KEY": "your-api-key"
}
}
}
}Once configured, your AI agent can use these tools:
Parse a web page:
"Parse https://example.com and give me a summary"
Parse a local PDF:
"Parse the file at /path/to/document.pdf as markdown"
Get a YouTube transcript:
"Get the transcript of https://www.youtube.com/watch?v=dQw4w9WgXcQ"
Parse with specific output format:
"Parse https://example.com as markdown"
# Install dependencies
npm install
# Build
npm run build
# Run locally
node dist/index.jsParseJet — Official website
Documentation — API reference and guides
Get API Key — Free API key (300 credits/month)
Pricing — Plans and pricing
TypeScript SDK — npm install parsejet
Supported Formats — PDF, DOCX, YouTube, web pages, images, and 25+ more
MIT
Get the transcript of a YouTube video, optionally in a specific language.
Inputs: url (required), language.
Parse a local file (PDF, DOCX, image, etc.) by uploading it to ParseJet and extracting its content.
Inputs: file_path (required), output_format.
Parse any URL (web page, YouTube video, PDF link, etc.) and extract its content as text or markdown.
Inputs: url (required), output_format.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 3 tools
Tools mostly distinct: parse_file handles local files, parse_url handles any URL, get_youtube_transcript is specialized for YouTube transcripts. There is slight overlap between get_youtube_transcript and parse_url for YouTube videos, but descriptions clarify different intents.
Two tools follow a parse_* pattern, while get_youtube_transcript breaks consistency with a get_ prefix. This is a minor deviation but still understandable.
3 tools for a 'Universal File & URL Parser' feels slightly thin but not unreasonable. The scope is covered by the two parse tools, with the transcript tool as an additional feature.
Core functionality (parsing local files and URLs) is covered. Missing explicit tools for structured data parsing (e.g., JSON, XML) but the parse_file and parse_url tools claim broad support. Minor gaps but adequate for common tasks.
Any URL to clean markdown/JSON; PDF, XLSX, CSV, DOCX to text and tables. Free tier for agents.
PDF, DOCX and image URLs to clean, LLM-ready Markdown or text, with OCR for scanned pages.
Web tools for AI agents: URL to clean Markdown, site crawl, screenshots, PDF/Office to text, OCR.
Document conversion and OCR for AI agents: PDF, Office docs, images to text.
