diff --git a/voussoir.net/writing/inkscape_batch_png_pdf/inkscape_batch_png_pdf.md b/voussoir.net/writing/inkscape_batch_png_pdf/inkscape_batch_png_pdf.md
new file mode 100644
index 0000000..c366b0f
--- /dev/null
+++ b/voussoir.net/writing/inkscape_batch_png_pdf/inkscape_batch_png_pdf.md
@@ -0,0 +1,74 @@
+Batch png to pdf with Inkscape & PDFtk
+====================================================
+
+Today, I had 57 png images of QR codes, and needed to create a single pdf document where each page contained a single QR code with the file's name shown centered above the QR. From start to finish, without planning ahead, it took me 23 minutes, a pace which would have required me to do 1 image every 24 seconds if I were to do them by hand. Plus, any subsequent tweaks can be done in a flash.
+
+I used [Inkscape](https://inkscape.org/) and [PDFtk](https://www.pdflabs.com/tools/pdftk-the-pdf-toolkit/).
+
+1. **Create one svg manually**
+
+    1. Open Inkscape and adjust the document page size to your liking.
+    2. Add the image to the page. I chose to embed the image rather than linking, but in retrospect, linking would have been the correct choice.
+    3. Add the text box to the page. With the text selected, press CTRL+SHIFT+T to open the Text and Font menu, choose Center alignment, and click apply.
+    4. Arrange the image and label to your liking.
+    5. Save the document.
+
+2. **Generate svgs for all other images**
+
+    1. Open the template svg in a text editor and copy everything.
+    2. Create a Python file, and start by:
+
+        ```Python
+        TEMPLATE = '''
+        paste the svg file here
+        '''.strip()
+        ```
+    3. If you chose to embed the image, it will be stored as base64. In that case, delete the actual base64 text, but leave the `data:image/png;base64,` prefix. After the comma, add `{image_base64}`.
+
+        `xlink:href="data:image/png;base64,{image_base64}"`
+
+        If you chose to link the image, you'll find the file path. Replace it with `{image_path}`.
+
+        `xlink:href="{image_path}"`
+
+    4. Find the element that contains the label text, and replace the text with `{label_text}`.
+
+    5. The rest of the Python script was as follows:
+
+        ```Python
+        import base64
+        import glob
+
+        filenames = glob.glob('*.png')
+        for filename in filenames:
+            label_text = f'...' # do some transformations on the filename
+            data = open(filename, 'rb').read()
+            image_base64 = base64.b64encode(data).decode('ascii')
+            svg = TEMPLATE.format(image_base64=image_base64, label_text=label_text)
+            open(filename.replace('.png', '.svg'), 'w').write(svg)
+        ```
+
+        Using linking instead of embedding, this would have been more like:
+
+        ```Python
+        for filename in filenames:
+            label_text = f'...' # do some transformations on the filename
+            svg = TEMPLATE.format(image_path=filename, label_text=label_text)
+            open(filename.replace('.png', '.svg'), 'w').write(svg)
+        ```
+
+3. **Convert svgs to pdf**
+
+    On Windows, a for-loop in the command line looks like this:
+
+    `for %x in (*.png) do inkscape --without-gui --export-pdf=%~nx.pdf %~nx.svg`
+
+4. **Combine all pages into single pdf**
+
+    `pdftk *.pdf cat output "all.pdf"`
+
+If I need to make any updates to the template, I can
+
+1. Open any of the svgs and make the changes.
+2. Copy it back into the Python file and redo the `{templating}`.
+3. Run the python script, inkscape export loop, and pdftk merger.