ePublisher AutoMap is the command-line publishing tool for WebWorks ePublisher. It runs the same builds as Designer unattended, from Windows Task Scheduler, GitHub Actions, Jenkins, or any CI/CD runner on Windows, and deploys the output to a folder or to Amazon S3.
AutoMap runs an ePublisher job, a Designer (.wep) or Express (.wrp) project, or a
composition job from the command line. It builds the targets with the project’s Stationery and deploys them to
a folder or to Amazon S3. --skip-reports speeds up large builds, and
--dryrun shows what a deployment would change without making it.
A build machine needs AutoMap, plus a licensed copy of Word or FrameMaker if the project publishes those sources, because ePublisher runs them during the build. DITA projects need Java 17 or later.
jobs:
publish:
runs-on: [self-hosted, windows]
steps:
- uses: actions/checkout@v4
- name: Publish the help
run: >
& "C:\Program Files\WebWorks\ePublisher AutoMap\WebWorks.Automap.exe"
--skip-reports --destination=docs-site docs\product.wep AutoMap is what makes "AI-ready output as a byproduct of every publish" practical without a human in the publishing seat. Scheduled and CI/CD-driven runs republish the help and refresh the Knowledge Base archive on the cadence the team already uses.
AutoMap's CLI shape — explicit project file, target name, output directory, optional flags — is operable by automation and by AI agents alike. The publishing pipeline becomes part of the AI-driven authoring loop.
AutoMap drives both Designer projects (.wep) and Express projects (.wrp) — the same engine, the same Stationery, the same output set. Recent additions include --skip-reports for faster CI/CD builds on large projects.
Works with nearly any version control or content management system. Pre-generation and post-generation scripts handle source retrieval, asset staging, and post-build deployment.
AutoMap's strategic positioning beyond CI/CD is as the publishing-pipeline surface that AI agents can natively operate. Agents draft Markdown++; agents publish via AutoMap; the resulting Reverb 2.0 helpset and Knowledge Files refresh on agent-driven cadence. No GUI in the loop; no human in the seat; the output is the same build a human triggers from Designer.
Documentation teams that author in Markdown++ have a source format that AI agents can read, interpret, and modify directly. Plain text in, plain text out — no XML overhead, no proprietary binary format.
An agent that has authored or modified a Markdown++ source can invoke AutoMap to publish it — running the same pipeline a human triggers from Designer. Explicit project file, target name, output directory, optional flags.
A WebWorks Claude Code skill pack — public on GitHub — teaches an AI agent how to traverse ePublisher projects, run AutoMap pipelines, customize outputs, and apply WebWorks conventions. Ready-to-use AI tooling for ePublisher today.
quadralay/webworks-agent-skillsGenerate all or selected output targets from a single AutoMap invocation — Reverb 2.0 helpset, Knowledge Base archive, PDF, ePUB, Eclipse Help, and the rest of the project Stationery configures.
Built-in script editor with environment variables for pre-generation source retrieval and post-generation deployment — the publishing pipeline integrates cleanly with the build environment around it.
A complete command-line interface drives every operation. AutoMap runs on developers’ machines, in GitHub Actions, in Jenkins, and in any CI/CD environment that can invoke a Windows binary.
Scheduled builds, CI/CD-driven publishes, and AI-agent-driven workflows all use the same AutoMap CLI. Talk to us about pricing for your team.
Contact WebWorks