From 13f5950389446a45932d6fff5ed141c1909ea3a1 Mon Sep 17 00:00:00 2001
From: "sergey.flakon" <sergey.flakon@mathpix.com>
Date: Mon, 6 Jul 2026 15:45:14 -0400
Subject: [PATCH] Add PDF -> XLSX conversion support
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

`mpx convert input.pdf output.xlsx` now works. xlsx is a supported
v3/pdf conversion output (extracts tabular content, one worksheet per
detected table). The PDF path downloads v3/pdf/{id}.xlsx directly, so
this only needs the new destination extension + dispatch case.

Scope: PDF sources only. Image and Markdown sources are intentionally
not wired up — those paths convert the OCR'd MMD via the /v1/export/*
endpoints, and there is no /v1/export/xlsx endpoint. Supporting them
would require backend work (a /v1/export/xlsx endpoint or rerouting
those paths through v3/converter) and is left as a follow-up.

Validated against the live API: POST /v3/pdf-file (file only) then
GET /v3/pdf/{id}.xlsx returns a valid .xlsx (HTTP 200), matching the
CLI's exact call pattern.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---
 README.md      | 6 +++++-
 lib/convert.js | 5 +++++
 lib/mpx.js     | 7 +++++--
 3 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/README.md b/README.md
index 58812c5..7370f03 100644
--- a/README.md
+++ b/README.md
@@ -31,15 +31,19 @@ mpx set-api-key ...
 # C:\Users\USERNAME\.mpx\config on Windows
 ```
 
-To digitize PDF's to editable Mathpix Markdown, docx, html or tex.zip:
+To digitize PDF's to editable Mathpix Markdown, docx, xlsx, html or tex.zip:
 
 ```
 mpx convert input-file.pdf output-file.mmd
 mpx convert input-file.pdf output-file.docx
+mpx convert input-file.pdf output-file.xlsx
 mpx convert input-file.pdf output-file.tex
 mpx convert input-file.pdf output-file.html
 ```
 
+`xlsx` extracts tabular content (one worksheet per detected table) and is
+currently supported for PDF sources only.
+
 To digitize images to editable Mathpix Markdown, docx, html or tex.zip:
 
 ```
diff --git a/lib/convert.js b/lib/convert.js
index ed0c324..76d63f9 100644
--- a/lib/convert.js
+++ b/lib/convert.js
@@ -432,6 +432,10 @@ async function PDFToDOCX(source, destination) {
   await ConvertPDF("docx", source, destination);
 }
 
+async function PDFToXLSX(source, destination) {
+  await ConvertPDF("xlsx", source, destination);
+}
+
 async function PDFToTEX(source, destination) {
   await ConvertPDF("tex", source, `${destination}.zip`);
 }
@@ -636,6 +640,7 @@ module.exports = {
   MarkdownToMarkdown,
   PDFToMMD,
   PDFToDOCX,
+  PDFToXLSX,
   PDFToTEX,
   PDFToHTML,
   ImageToMMD,
diff --git a/lib/mpx.js b/lib/mpx.js
index 2c7808f..66c55e1 100755
--- a/lib/mpx.js
+++ b/lib/mpx.js
@@ -103,7 +103,7 @@ program
 
 program
   .command("convert <source.ext> <destination.ext>")
-  .description("convert files between markdown, mathpix markdown, docx, latex and pdf formats")
+  .description("convert files between markdown, mathpix markdown, docx, xlsx, latex and pdf formats")
   .addOption(
     new commander.Option("-p, --pdf-method <type>", "choose to make pdf from latex or html")
       .choices(["latex", "html"])
@@ -133,7 +133,7 @@ program
         return process.exit(1);
       }
 
-      const validDestinationExtensions = [".mmd", ".md", ".pdf", ".tex", ".html", ".docx"];
+      const validDestinationExtensions = [".mmd", ".md", ".pdf", ".tex", ".html", ".docx", ".xlsx"];
       if (!validDestinationExtensions.includes(destinationExt)) {
         console.log(
           `Invalid destination extension, must be one of: ${JSON.stringify(
@@ -163,6 +163,9 @@ program
         case sourceExt === ".pdf" && destinationExt === ".docx":
           await convert.PDFToDOCX(source, destination);
           break;
+        case sourceExt === ".pdf" && destinationExt === ".xlsx":
+          await convert.PDFToXLSX(source, destination);
+          break;
         case sourceExt === ".pdf" && destinationExt === ".tex":
           await convert.PDFToTEX(source, destination);
           break;