PDF Munk

@pdfmunk parse the Q4 financial report at https://example.com/q4-report.pdf and extract all tables with their data

I've extracted 7 tables from the Q4 report. Table 1 (page 3) contains quarterly revenue by region: North America $4.2M, EMEA $2.8M, APAC $1.9M. Table 2 (page 5) shows operating expenses broken down by department. Tables 3-5 cover customer acquisition metrics, and Tables 6-7 detail cash flow statements. Each table includes row and column headers with all numeric values preserved. The full structured data is available for further analysis or export.

This example demonstrates PDF Munk's structured parsing capability, which extracts tables with their semantic layout intact. The PDF must be publicly accessible via URL — private or authenticated documents require pre-upload to a public endpoint. Large PDFs may hit rate limits on free-tier API keys.

@pdfmunk generate a PDF from https://github.com/ourteam/project/wiki/Architecture and merge it with our existing design-spec.pdf to create a combined reference doc

I've generated a PDF from the Architecture wiki page (8 pages, includes all diagrams and code blocks) and merged it with design-spec.pdf. The combined document is 23 pages total and temporarily hosted at https://pdfmunk.com/tmp/arch-spec-merged-a8f3.pdf. This URL expires in 24 hours. The merge preserved all formatting, though the wiki's syntax-highlighted code blocks rendered as static images in the PDF.

This showcases PDF Munk's generation-plus-merge workflow, useful for assembling documentation from mixed sources. Both operations consume API quota. The temporary hosting means you should download or re-host the result if you need it beyond 24 hours. HTML-to-PDF rendering may not perfectly match browser display for complex CSS.

@pdfmunk OCR the scanned invoice at https://example.com/invoice-scan.pdf and tell me the total amount due and payment terms

I've OCR'd the invoice and extracted the following: Total Amount Due is $3,847.52 (line 47 of the OCR output). Payment Terms state 'Net 30 from invoice date of March 15, 2024' (line 52), making the due date April 14, 2024. The invoice also lists three line items: consulting services ($2,800), travel expenses ($847.52), and a 10% early-payment discount available if paid by March 25. OCR confidence was high (95%+) for all numeric fields.

This example pairs PDF Munk's OCR tool with AI reasoning to answer a specific question from a scanned document. OCR accuracy depends on scan quality — low-resolution or skewed images may produce garbled text. The MCP returns line-by-line results, which the AI then interprets. API key must have OCR quota enabled.

When PDF Munk beats manual proposal builds

A 6-person consulting shop sends 15-20 custom proposals a month. Each proposal pulls boilerplate from Google Docs, case studies from Notion, and pricing tables from Airtable. Before PDF Munk, someone spent 90 minutes per proposal copying content into InDesign. Now the team uses Generate PDF to render HTML templates directly and Merge PDFs to stitch in standard appendices. The workflow drops to 15 minutes. The threshold: if your proposals need pixel-perfect brand control or complex typography, you still need a designer. But if speed and consistency matter more than custom layout, PDF Munk wins. Set up the integration once and let anyone on the team ship proposals without waiting on design.

Why OCR PDF works for small-team bookkeeping

A 3-person agency processes 40 vendor invoices a month. Invoices arrive as scanned PDFs in email, and someone manually types line items into QuickBooks. The team adds PDF Munk's OCR PDF tool to extract text, then pipes the results into a Switchy prompt that formats entries for their accounting system. The workflow cuts data entry from 20 minutes per invoice to under 2 minutes. The catch: OCR accuracy depends on scan quality. If your vendors send clean digital PDFs, Parse PDF is faster and more reliable. But for scanned or image-based invoices, OCR PDF handles the job without a third-party data entry service. Worth testing on your worst 5 invoices before committing.

When Convert Images to PDF saves support handoffs

A 10-person SaaS support team triages 200 tickets a week. Customers attach 3-5 screenshots per bug report, and engineers need those images in a single reference doc. The team uses Convert Images to PDF to bundle screenshots into one temporary PDF, then shares the link in Linear. The PDF expires in 24 hours, which matches the ticket lifecycle. The alternative is asking customers to zip files or manually stitching screenshots in Notion. The trade-off: if your team needs permanent archives or searchable image text, this MCP isn't the answer. But for fast triage where context expires quickly, bundling images into a throwaway PDF keeps the handoff clean and the ticket moving.