When errors occur
Errors can surface:
During AI generation (invalid/missing values, GPT issues).
When pushing to Magento (required fields, uniqueness, URL key conflicts).
From store configuration problems (e.g., URL suffixes, disabled attributes).
How to spot errors
Open Apps → Content enrichment → Lists, choose your job.
Items with issues are flagged. Open an item to see the detailed message.
For GPT-related problems, you’ll also see a job‑level notification indicating why generation failed.
Common Magento-side errors and fixes
Error type | Cause | How to fix |
Missing required attribute | A required Magento field is empty. | Fill manually or Regenerate with AI. |
Invalid attribute value | Value doesn’t match allowed options. | Correct the value or Transform to an existing option. |
URL key already exists | URL key not unique. | Edit to a unique URL key. |
Attribute must be unique | E.g., EAN or SKU already used. | Provide a unique value. |
Store connection error | API credentials/connection issue. | Check Settings → Connection settings. |
Preview URL not working | Missing | Add suffix: General settings → Settings → Connection settings → Edit store → set category/product suffix → Save. |
AI (GPT) specific errors
Where you’ll see them:
Job banner/notification explaining the failure (e.g., token limit, rate limit, auth).
Item level when you click Regenerate with AI.
Typical causes and resolutions:
GPT error | What it means | How to fix (Elovate) | How to confirm (provider) |
Empty tokens / out of credits | Your OpenAI balance is depleted. | Pause/stop generation, then resume after funding. Optionally Regenerate failed items. | Check OpenAI → Billing. Ensure Pay‑As‑You‑Go is active and funded. |
Invalid API key (401) | Key missing/incorrect/revoked. | Re‑enter key in Settings → Integrations → OpenAI. | Verify the API key in OpenAI → API Keys. |
Rate limited (429) | Too many requests too quickly. | Retry later, reduce batch size, stagger regenerations. | Check usage/rate‑limit notices in OpenAI dashboard. |
Model/endpoint unavailable | Temporary model outage or region issue. | Retry; if persistent, switch to the default supported model. | Check OpenAI status page / dashboard. |
Prompt/content length | Inputs exceed model limits. | Trim “Use these attributes…” inputs or reduce “Maximum characters”; simplify Additional prompt. | N/A—adjust and retry. |
Reproducing/clearing a GPT error
To reproduce an item error for testing, open the item and click Regenerate with AI.
To clear after fixing the cause (e.g., added funds), use Regenerate at list level (reruns the job) or re‑prompt at item level.
Tip: If many items failed due to the same GPT cause (e.g., credits), first resolve the cause, then use list‑level Regenerate instead of fixing item‑by‑item.
Tools for fixing errors in Elovate
Skip element – Completes the item without changes (removes from list).
Transform value – Map an invalid value to a valid existing option.
Add new option – Create a new attribute option directly from the error.
Regenerate with AI – Rerun AI for a single item (or use Regenerate at list level).
Editor and moderator permissions
Editors only: can fix and save; can push to Magento only if no moderators are assigned.
Editors + moderators: only moderators can approve and push (bulk approve available in My moderation tasks).
Same user is editor + moderator: full capabilities, including bulk approval.
Best practices
Fix validation errors before approval to prevent sync failures.
Keep AI prompts specific; include structure rules (headings, bullet limits).
If GPT errors occur, check the job banner, then verify OpenAI billing/API key and retry.
Use list‑level Regenerate after resolving a systemic GPT issue; use item‑level re‑prompt for one‑offs.