Word, PowerPoint & Excel to Markdown — for RAG & AI Agents
Pricing
from $10.00 / 1,000 document converteds
Word, PowerPoint & Excel to Markdown — for RAG & AI Agents
Convert DOCX, PPTX, and XLSX files to clean, LLM-ready markdown at scale. Headings, tables, slides, and sheets preserved. Never-fail runs, per-document output. The Office twin of PDF-to-Markdown.
Pricing
from $10.00 / 1,000 document converteds
Rating
0.0
(0)
Developer
Shawn Downs
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
2 days ago
Last modified
Categories
Share
Word, PowerPoint & Excel to Markdown
Convert DOCX, PPTX, and XLSX files into clean, LLM-ready markdown at scale — headings, tables, slides, and sheets preserved. The Office twin of our PDF-to-Markdown actor: point agents and RAG pipelines at Office documents the same way they already handle PDFs. No subscription, no server, pay per document.
Convert Word / PowerPoint / Excel for RAG
Feed a list of document URLs (.docx, .pptx, .xlsx), get markdown back. Headings become
markdown headings, tables become GitHub-flavored markdown tables, slides and sheets are
labeled. Use perSectionRows: true to receive one row per heading section / slide / sheet —
ready to chunk straight into a vector store.
Output
Per document: url, type (docx|pptx|xlsx), markdown, sections[], char_count. With
perSectionRows, one row per chunk: url, type, section, markdown. Corrupt,
password-protected, or non-Office files come back as error rows ({url, error}) — the run
itself never fails, and you only pay for documents actually converted.
Why this actor
- Built on Microsoft's open-source MarkItDown (MIT), with a mammoth fallback for Word heading fidelity — no proprietary black box.
- Pure file parsing: no scraping, no anti-bot, nothing external to break.
- Never-fail batch runs with per-document error rows.
- Pairs with PDF-to-Markdown for full document coverage in one pipeline.
Use with AI agents
MCP-friendly: hand an agent a Word/PowerPoint/Excel URL and get structured markdown back in one call, paying cents per document. Crawl → collect document links → convert here → embed.
