Apify Document Tools MCP Server

Apify document tools

Open-source Apify Actors for working with business documents, audio, video, websites and domains. Run them in the cloud with no setup from Apify Store, call them from your code, or plug them into AI agents through the Apify MCP server.

Ready-made examples (Apify tasks): document-to-markdown examples, office-to-pdf examples, OCR examples — every Actor page has an Examples tab.

Use from AI agents (MCP)

All Actors are available as MCP tools — hosted (no install): https://mcp.apify.com?tools=ai_admin/document-to-markdown,ai_admin/office-to-pdf,ai_admin/pdf-image-converter,ai_admin/audio-transcriber,ai_admin/image-to-text-ocr,ai_admin/text-to-speech,ai_admin/website-screenshot,ai_admin/domain-whois-dns-ssl,ai_admin/video-frame-extractor — or as a local stdio server in mcp-server/.

Built with

LibreOffice, Tesseract OCR, pdfplumber, pypdfium2, markitdown, openpyxl, pypdf, Piper TTS, Playwright, FFmpeg and the Apify Python SDK.

Development

Each Actor folder is self-contained (Dockerfile, .actor/, src/, tests/):

cd document-to-markdown
python tests/make_corpus.py /tmp/corpus          # 32 sample documents with known content (needs LibreOffice)
python -m tests.check local /tmp/corpus          # convert locally and compare with the expected content
python tools/deploy.py                           # create/update the Actor on Apify (APIFY_TOKEN) and build it

cd ../office-to-pdf
python -m tests.check local /tmp/corpus --pdfa   # Office -> PDF/A on the same corpus

cd ../audio-transcriber
GROQ_API_KEY=... python -m tests.check /path/to/media   # real Groq calls; see tests/check.py for the sample files

cd ../pdf-image-converter
python -m tests.check /tmp/corpus                # PDF <-> image conversions

cd ../image-to-text-ocr
python -m tests.check /tmp/corpus                # OCR of images, scans, TIFF, HEIC, rotated pages

cd ../text-to-speech
GROQ_API_KEY=... python -m tests.check /tmp/tts /tmp/voices   # speech is checked by transcribing it back

cd ../website-screenshot
python -m tests.check                            # screenshots and PDFs of a local test site (needs Chromium)

cd ../domain-whois-dns-ssl
python -m tests.check                            # RDAP and DNS; WHOIS and SSL are checked on Apify (tools/test_on_apify.py)

cd ../video-frame-extractor
python -m tests.check /tmp/video                 # frame modes on a generated 4-scene video (needs FFmpeg)

Local runs need Tesseract (eng, ukr, rus), LibreOffice, FFmpeg, Chromium and the packages from requirements.txt.

Bug reports with a sample document (or a similar public file) are welcome as GitHub issues.

License

MIT. Text to Speech uses the Piper engine (GPL-3.0) as a separate package.