PitchAI

Move a publication without losing its history

PitchAI built a populated editorial platform with an explicit record of where existing content went, how readers would reach it and how the site could be restored.

Knowledge publishing
Deployed staging platform; production cutover separate
Representative migration ledger: an existing article is imported, an older duplicate is mapped to a canonical page, and a test page is explicitly skipped. Each decision retains its source reference.
Representative migration decisions using synthetic labels. No customer content, source addresses or editorial accounts are shown.

A new editor still needs the old publication

The brief called for a custom website and content management system, with media handling and automated, downloadable backups. Existing articles and pages also needed a place in the replacement. A working editor alone would leave that migration unfinished.

PitchAI built a Wagtail platform with structured page content, author profiles, categories, related reading and draft review. Headings can form an article navigation, while documents and media remain editable parts of the publication.

Account for each source record

The import works from a preserved source snapshot. It records each source page as imported, mapped to another destination or deliberately skipped. A duplicate can point to its canonical page; a known test page can be excluded without disappearing from the accounting.

Each ledger entry retains its source reference, content fingerprint and reconciliation note. If a previously recorded source changes, the importer raises an error. Editors can inspect the migration map instead of relying on a claim that everything was copied.

Carry the addresses forward too

The migration creates permanent redirects for old content, category, author, attachment and file addresses, avoiding redirects back to themselves. Imported publication dates are retained, and links to migrated pages and assets are rewritten where a destination exists.

A route inventory gives validation an explicit list to check. Delivery records cover the populated site and its legacy redirects, alongside browser checks of editorial review, media, documents and protected account surfaces. Normal startup skips the initial import once the content site exists, preserving subsequent editorial work.

Test the return journey

Backups were already part of the brief. The implementation makes recovery inspectable: portable archives include content and media, with fingerprints that can be checked before restoration. The restore process requires an isolated, fresh destination rather than overwriting an existing editorial database.

The deployed staging release was also restored into an isolated environment. Its content, assets, migration records and routes were checked there, then the temporary restore environment was removed. That provides evidence of a usable recovery path beyond the presence of a download button.

The result is a staging publication that editors can review with its migration and recovery evidence alongside it. The customer's original production website remained separate; this work is not presented as a completed production cutover.

Browse all case studies