✦ Free & local-first · 4 servers · 32 tools · zero signup

One MCP suite for
media, web data, email & documents

Everything your AI agent touches daily — one-line install, runs on your machine, no API keys. Optional cloud tier for heavy lifting.

Try in 10 seconds → PyPI MCP Registry
32agent tools
4servers
0API keys needed
100%local & private

MediaMCP

Convert, compress, thumbnail any video/audio with ffmpeg. Metadata & clean transcripts from YouTube and 1000+ sites. Optional cloud Whisper transcription.

$uvx --from mcpproducts-media media-mcp
media_infoconvertcompressthumbnailvideo_infotranscriptcloud_transcribe ✦
PyPIMCPize
media-mcp — illustrative example

ExtractMCP

Turn any URL or raw HTML into clean data: readable markdown, tables as JSON, resolved links, JSON API reads. The "Firecrawl without an API key" layer.

$uvx --from mcpproducts-extract extract-mcp
fetch_markdownextract_textextract_tablesextract_linksextract_jsoncloud_extract ✦
PyPIMCPize
extract-mcp — illustrative example

OfficeMCP

The document workhorse: read & create PDF, Excel, Word, PowerPoint; convert csv ↔ xlsx ↔ json. Pure Python, works everywhere.

$uvx --from mcpproducts-office office-mcp
pdf_textpdf_from_textxlsx_read/writedocx_createpptx_outlineconvert_tabularcloud_ocr ✦
PyPIMCPize
office-mcp — illustrative example

MailMCP

Read, search and draft email over standard IMAP — Gmail, Outlook, any server. Never sends: drafts land in your Drafts folder, you press Send.

$uvx --from mcpproducts-mail mail-mcp
list_recentsearch_emailsread_emaildraft_emailmove_emailcloud_digest ✦
PyPIMCPize
mail-mcp — illustrative example

Why agents pick these

🏠

Local-first

Free tools run on your machine — your files and email never leave it. No accounts, no keys, no rate limits.

⚡

One line each

uvx --from mcpproducts-media media-mcp and you're running. Works with Claude Desktop, Claude Code, Cursor, any MCP client.

🔐

Safe by design

Mail drafts but never sends. No telemetry. Open package on PyPI, listed in the official MCP Registry.

☁️

Cloud when you need it

Whisper transcription, JS-page rendering and OCR available as an optional keyed tier — free beta while we scale.