Автоматизированный парсер и YML-генератор для товара Google Sheets → Supabase → YML для OpenCart на n8n

Goals and brief overview of the task:

It is necessary to fully automate the collection, enrichment, and preparation of product data for an online store (or marketplace on OpenCart). The source of products is Google Sheets, the database is Supabase. Then, the products undergo AI processing (names, descriptions, SEO, multilingualism), photo filtering via Google Vision, and are automatically exported in YML format for import through UniXML into OpenCart (according to the Rozetka template).

All tasks are implemented as understandable automated scenarios in n8n (cloud/self-hosted).

Automation stages:

1. Import from Google Sheets into Supabase:

   • Product data is automatically transferred from Google Sheets to Supabase.

   • Scheduling/triggers are possible (via button or timer).

2. AI processing (OpenAI / BrowseAI):

   • Generation or optimization of names, descriptions, SEO blocks.

   • Translation UA/RU.

   • Automatic attribute filling if there are gaps.

3. Validation and photo filtering via Google Vision:

   • Check minimum number of photos (min. 2, recommended 3-5).

   • Filter out photos with “watermarks,” poor quality, duplicates.

   • Mark problematic/missing photos in logs.

4. YML/XML generation according to Rozetka template (UniXML):

   • Forming a file ready for the UniXML module in OpenCart.

   • Automatic creation and updating of YML, so the catalog can be uploaded/updated on the marketplace at any time.

5. Logging and monitoring:

   • All “empty” fields, duplicates, missing photos, errors are recorded in a dedicated Google Sheets for quick review.

Developer requirements:

• Experience with n8n (preferably — AI/API: OpenAI, BrowseAI, Google Vision, Supabase, Google Sheets).

• Understanding of YML file structure for marketplaces/UniXML.

• Ability to automate scenarios with minimal manual intervention.

Additional desirable skills:

• Optimization ideas (e.g., smart category/brand mapping, filters, integration with additional sources, auto-backup, error notifications, etc.).

• Scaling options: support for multiple stores, multiple sources (API/PDF/others).

Main tasks (core technical assignment):

1. Connect Google Sheets and Supabase (set up two-way synchronization or scheduled updates).

2. Configure automation chain in n8n:

   • Sheets → Supabase → AI (OpenAI/BrowseAI) → Google Vision → YML/XML generator → UniXML.

3. Implement multilingual support (minimum UA, RU).

4. Configure photo processing and validation via Vision (number of photos, quality, watermark filtering).

5. Generate YML exactly according to Rozetka template (link), so that the file is compatible with UniXML.

6. Logging and convenient monitoring (Google Sheets, email or Telegram notifications).

Frequently asked questions:

• What are the API keys?
  — All keys (Google, OpenAI, Supabase, BrowseAI) will be provided after selecting the contractor.

• Is it necessary to work with categories/attributes?
  — Yes, it is desirable to have automatic mapping/unification.

• What is the minimum number of photos per product?
  — Minimum 2 photos (recommended 3-5).

• Is translation/generation needed for two languages?
  — Yes, Ukrainian + Russian.

• What YML template should be used?
  — Use UniXML (Rozetka).

• What is the scope of the task?
  — Approximately up to 10,000 products, with selective categories.

Ready to hear your ideas:

Add in your response:

• How would you build this chain in n8n?

• Which nodes/integrations would you use?

• What additional features do you propose?

• Examples/demos of your similar automations?

• Estimated implementation timelines.

Clarification for the contractor:

• Alternative solutions can be proposed (e.g., Apify, Integromat, Zapier, custom Python scripts if they are easily run from n8n).

• Main focus — automation and ease of management, clear logging, and support for changes.

Questions that may arise for contractors (FAQ):

1. What is the preferred tech stack?

• Main: n8n, API Supabase, BrowseAI/OpenAI, Google Vision, Google Sheets.

2. How often should the parser run?

• On schedule (e.g., once a day), or manually triggered.

3. What services are already available (keys/accounts)?

• Supabase is already created, keys will be provided.

• BrowseAI, OpenAI – under discussion (testing desired).

4. Is a web admin panel needed?

• Optionally, can be replaced with Google Sheets/spreadsheet.

5. Should API limits be considered?

• Yes, it is recommended to implement delays or error handling (retry/logging).

6. What YML/XML format is required?

• For the marketplace (example provided), preferably — a ready link or Google Drive.

7. Is integration with OpenCart needed?

• It is welcome, but YML for import is sufficient for now.

Provide your price and timeline.

Be realistic — interested in long-term cooperation on other projects.