AI Content Migrate

Automate content modeling and migration from legacy websites and internal tools into Drupal.
The module analyzes the source, proposes content types and fields,
imports media and taxonomies through generated
Migrate API migrations, and creates Canvas templates for Drupal CMS.
AI is used for high-level analysis only; every proposal is reviewed by a person before anything is imported.

Modes

• AI agent mode: an AI Content Migrate agent analyzes individual pages and proposes pages and fields. Suited to simple or fragmented migrations.
• Sitemap (Drupal CMS) mode: a guided workflow starting from a sitemap. It groups pages into content types, lets you refine the model visually, generates Migrate API files (File → Media → Node) and builds a Canvas template per content type. Translations are detected from hreflang attributes.
• Database mode: migrates internal tools with no public frontend (e.g. Confluence, BookStack) from a database dump. A source profile recognizes the application from its schema and migrates spaces, the page tree, labels, attachments and authors, with URL aliases and redirects from the legacy URLs. For an unknown application, the AI drafts a profile from the database structure only; you verify and approve it, and later dumps are recognized without any AI call.

Requirements

• AI with a configured chat provider, and AI Agents
• Migrate Plus and Inline Entity Form
• chrome-php/chrome and a headless Chromium binary, for JavaScript-rendered pages
• Optional: Canvas (layouts), Pathauto and Redirect (database mode URLs and redirects)

Key features

• AI-assisted content models: content types, fields, taxonomies and media, reusing existing structures where possible.
• Native Migrate API output: standard migrate_plus YAML with dependencies, runnable from the UI (batch, rollback) or Drush.
• Field ↔ selector mappings for web pages (e.g., field_title → //h1), with safe fallbacks.
• Canvas templates generated and validated against Canvas' own rules, fully editable in the Canvas editor.
• Database mode: source profiles (Confluence built in), an AI-assisted profile builder with local verifit from HTML exports of the legacy pages.
• Human in the loop at every step: review, adjust and approve before importing.

How it works

1. Choose a mode: an agent, a sitemap, or a database dump (optionally with the application's files and HT
2. The module analyzes the source and proposes a content model and mappings.
3. You review and adjust them, then generate the migrations.
4. Run the migrations, check the result, and roll back if needed.

Installation

1. Install the module and its dependencies with Composer, and install headless Chromium on the server.Configure an AI provider in the AI module.
2. Enable AI Content Migrate.
3. For the database mode, configure the private file system so uploaded dumps and files stay private.
4. Start from Configuration → AI: the agent, the sitemap import or the database migration.

Notes

• Always review AI suggestions before approving them.
• In agent and sitemap mode, HTML of the analyzed pages is sent to your AI provider; check your complian
• In database mode, no data of the dump is sent to the AI: only its structure, statistics and codes (e.g. a content type or status value), and a text-free outline of the HTML exports.
• The module is experimental: validate the generated models and migrations before running them on production