# How LaunchLayer Works

Canonical page: https://launch-layers.cloud/help/launchlayer-end-to-end
Last updated: 2026-10-02

## 1. Start with the authoritative input
Create a project from what you actually control: a live HTTPS web app, GitHub repository, source ZIP, existing signed IPA/APK/AAB, or an existing store application.

## 2. Inspect before changing anything
LaunchLayer records source evidence, app identity, framework/build hints, versions, assets, credentials, and provider state. It uses that evidence to determine the release path instead of assuming every app is the same.

## 3. Configure the supported native path
For supported wrapper projects, LaunchLayer prepares Capacitor-oriented configuration, native capability requirements, store metadata guidance, permission copy, dependency/setup instructions, and release checklists.

## 4. Prepare signing and store connections
Apple and Google releases require the correct developer accounts, app identifiers, signing material, store records, and permissions. External providers remain independent systems with their own rules.

## 5. Build using the path the project supports
Legacy hosted-wrapper cloud builds use generated GitHub repositories and GitHub Actions. Existing signed binaries can move through inspection and supported direct-upload flows. Arbitrary GitHub/ZIP Source Build V2 compilation remains planned.

## 6. Validate before upload
LaunchLayer checks release identity, signing/readiness evidence, required assets and metadata, and known mismatch conditions. Passing those checks does not guarantee provider approval.

## 7. Upload and monitor
Supported deployment flows upload an already-built, correctly signed artifact to the matching store application, preserve job history, and surface provider processing state. Apple and Google control review and publication.

## 8. Repeat for every release
A new version normally means re-inspecting changed source or a new binary, updating version/build identity, rebuilding when native inputs changed, validating again, and uploading the new artifact.

## AI assistance
Each project includes a project-scoped AI Agent for evidence-backed questions and guarded local-setting proposals. The model does not receive unrestricted production mutation authority.