Code and data export

What you get

• Every app's code as a Git bundle
• Every database as a standard PostgreSQL dump
• Members, groups and apps listed as JSON
• Secret names, never their values
• The full audit log
• Every whisk.yaml each app has used
• Available on every plan, in every account state

How it works

An owner or admin starts an export from the Trust page in the dashboard, or with whisk export. Whisk writes one compressed archive into the business's storage bucket. A README inside lists the commands that bring each part back.

Everything comes in open formats: code as a Git bundle, and each database as a PostgreSQL dump that loads into any Postgres server. The CLI, the whisk.yaml format, the agent skill and the templates are open source (MIT). Each app is an ordinary container that follows plain rules.

Owners are emailed when the export is ready, and the archive is kept for seven days. Each download link is made when an owner or admin asks for it and works for five minutes, so someone who loses access cannot keep downloading. Export works even when an account is frozen for an unpaid bill, because leaving should not depend on being in good standing. To keep a live copy of the code elsewhere, use the GitHub copy.

Example

An office furniture dealer keeps a quarterly export of its quoting and stock apps in its own document archive. Its auditor can see that the business holds a complete copy of its code and data, readable without Whisk.

For your coding agent

whisk export --wait
git clone repo.bundle <app>
pg_restore --no-owner -d "$DATABASE_URL" production.dump

--wait follows the export and prints the download link when it is ready; whisk export show <id> prints a fresh one later. Every flag is in the CLI reference.