Skip to main content
Pandoc conversions are run by the clsi service of every Ayakaleaf Pro pod, like compiles: clsi asks the Docker Engine of the node to start a container from PANDOC_IMAGE. The web service shows the import and export menus. What the feature does and how to build the image is described in Pandoc Import and Export.

Configuration

1

Put the image into Docker on every node

Build the image once from the develop/pandoc folder of the ayakaleaf-pro repository, then load it into Docker on each node that can run Ayakaleaf Pro:
Or push it to a registry the nodes can reach and sudo docker pull it on each node. As with the TeX Live images, it must be in Docker, not only in containerd.
2

ConfigMap

Both web and clsi read ENABLE_PANDOC_CONVERSIONS from the same ConfigMap, so they cannot disagree on Kubernetes.
3

Apply and restart

Several replicas

Nothing extra beyond having the image on every node: each replica converts on its own node.

Tested on Kubernetes

With the image built from develop/pandoc and loaded into Docker on every node with docker save | docker load, on Kubernetes 1.37 with one replica, our end-to-end tests for Pandoc pass (6 cases): importing a Markdown document as a LaTeX project, the error shown for a broken Word document, exporting Word (which imports again), Markdown and HTML, and the menus hidden without ENABLE_PANDOC_CONVERSIONS.
Last modified on November 2, 2026