Agentic app packaging for modern Windows delivery.
Use EtherApps Forge for agentic app packaging with Foundry-backed orchestration to capture applications in a controlled virtual machine, choose the right package route, and prepare MSIX, MSI, PowerShell App Deployment Toolkit, and Intune-ready outputs for review.
Your next application package starts with a conversation.
Connect an AI assistant to EtherApps Forge and ask it to inspect the capture,
help choose an output format and prepare the packaging steps.
Your engineer can review the work,
sign the package and run testing and validation through Forge.
Try the workflow on your own application.
Forge is free for seven days. No card needed.
VM-based
capture and package from a controlled Windows environment
4 outputs
MSIX, MSI, PSADT, and Intune-ready package paths
Reviewed
agentic workflow with human approval before release
The problem
Packaging stays slow when the decision work is still manual.
Packaging teams are asked to move faster while dealing with missing installers, unclear dependencies, and multiple deployment targets. Traditional tools still leave skilled packagers to decide the route, rebuild scripts, prepare Intune metadata, and document the decision manually.
Installer media is missing or unreliable
Many useful applications only exist as installed footprints on old machines. A VM-based capture route gives the team a practical starting point when clean source media is gone.
The output route is not obvious
MSIX, MSI, PSADT, and Intune-ready packages each fit different deployment paths. The agent helps recommend the right route before rework starts.
Packaging judgement is scarce
Experienced packagers still make the final call, but agentic triage can remove repetitive inspection, script preparation, and metadata work from the critical path.
Teams need control, not black-box release
The workflow keeps human review and evidence around the package route, generated artefacts, and release readiness before anything reaches users.
What changes
What changes
Package from a controlled VM
Start with the application as it runs, capture the footprint in a controlled Windows environment, and reduce dependency on perfect source installers.
Let the agent inspect before packaging
Use Foundry-backed orchestration to analyse application state, dependencies, and destination requirements before the package route is chosen.
Prepare the right output
Create MSIX, MSI, PowerShell App Deployment Toolkit, or Intune-ready package paths from the same capture-led workflow.
Keep reviewer sign-off
Give the packaging team a reviewable decision trail and release checkpoint, rather than asking an agent to publish without control.
Capture from running systems
Live-capture wizard: file, registry, AppData, services, and dependency capture from a running Windows system when installer media is missing or incomplete.
Forge's AI controller inspects the captured application state, recommends MSIX, MSI, PowerShell App Deployment Toolkit, or Intune-ready output, and prepares scripts and metadata. The packaging team keeps the final approval before release.
Walkthrough
See Forge packaging in action.
A focused walkthrough of VM capture, AI-assisted route selection, signing, manifest fix-ups, and Intune-ready outputs.
Feature deep dives
The packaging workflow, broken down.
VM-based capture, MSIX outputs, and MCP integration for external orchestration. One focused video per area so packaging, endpoint, and automation teams can jump to the part that matters.
1:12
Capture a running Windows application from a controlled VM: files, registry, AppData, services, and dependencies, without installer media.
1:08
Produce signed MSIX, AppAttach, IntuneWin, and MSI packages from one capture, ready for Intune, AVD, and Cloud PC delivery.
1:36
Drive Forge from your own AI orchestrator through the Forge MCP server: a tool-integration interface for external agents and pipelines. Separate from Forge's internal AI controller.
How we deliver it
Product mapping
This route is led by EtherApps Forge. Forge supplies the VM-based capture and packaging workflow; Microsoft Foundry-backed agentic orchestration helps inspect application state, choose the package route, generate scripts and metadata, and prepare outputs for review. Use the MSIX page when the buyer is specifically searching for MSIX, signing, or App Attach. Use application modernisation when the buyer is planning a broader migration programme.
Where this route fits best
Packaging backlogs where installers are missing and the installed application footprint is the only reliable source.
Intune migration projects that need repeatable package preparation, install commands, detection logic, and deployment metadata.
Windows 11, Azure Virtual Desktop, or Cloud PC readiness work where application blockers are delaying the rollout plan.
MSP packaging-factory operations that need a repeatable workflow across multiple customer estates.
Teams moving from manual packaging triage to a controlled AI-assisted packaging pipeline.
Capture installed Windows applications, reduce packaging unknowns, and produce modern deployment-ready outputs faster.
It means the packaging workflow uses an AI agent to inspect a controlled virtual machine capture, reason over application behaviour, recommend the package route, and prepare outputs for review. The package is still approved by a human before release.
How is this different from MSIX packaging?
MSIX packaging is a specific output route. Agentic app packaging is the wider orchestration layer that helps decide whether an app should become MSIX, MSI, PowerShell App Deployment Toolkit, or an Intune-ready package.
Does the agent publish packages without review?
No. The intended workflow keeps a reviewer checkpoint before release. The agent accelerates capture analysis, route selection, script preparation, and metadata work, while the packaging team keeps control.
Which outputs are in scope?
The route is built around MSIX, MSI, PowerShell App Deployment Toolkit, and Intune-ready outputs. MSIX App Attach and wider migration routes still connect back to the dedicated MSIX and application modernisation solution pages.
What happens if the AI-recommended package route turns out to be wrong?
The route recommendation is a starting point, not a commitment. Because the reviewer checkpoint sits before release, a packager can choose a different output, MSIX, MSI, PSADT, or Intune-ready, than the one the agent proposed, without redoing the VM capture. The capture is reusable across routes; only the packaging step downstream of it changes.
Does this replace the Win32 Content Prep Tool or Microsoft's own Intune packaging steps?
No. Forge produces the same kinds of outputs (.intunewin, MSIX, MSI) that Microsoft's own Win32 Content Prep Tool and MSIX tooling create, and those packages still upload into Intune the normal way. What Forge changes is the work before that upload: capturing the app from a controlled VM and using agentic orchestration to prepare the package and its metadata, instead of a packager doing that preparation by hand.
Is there an AI tool that packages legacy Win32 applications automatically?
Yes. Agentic app packaging captures a legacy Win32 application from a controlled VM, then an AI agent inspects that capture, reasons over the application's runtime behaviour, and recommends the package route, MSIX, MSI, PowerShell App Deployment Toolkit, or Intune-ready, before preparing the output and its metadata. A human reviewer still checks and approves the package before it ships, so "automatic" means less manual detective work per app, not an unattended release process.
Use a 7-day Forge trial on a real packaging blocker, or book a packaging fit check to review the VM capture, agentic route selection, package outputs, and approval model before a wider rollout.
MSIX at Scale: an automated packaging study of 3,261 Windows application cases. Preprint published on Zenodo, not yet peer reviewed. Applications came from catalogue searches, not a random sample. A smoke test is a first check, not proof of production readiness. EfficientEther builds EtherApps Forge and declares that interest. Read the published study
We use cookies on this site
We use strictly-necessary cookies to make the site work. With your consent we also use analytics cookies (Google Analytics and Microsoft Clarity) to understand how visitors use the site so we can improve it. We do not run advertising cookies. Cookie policy·Privacy policy