Using Claude (AI) to Run Bulk Imports in Luminous

Claude can now trigger any import that Luminous supports — directly through a conversation — allowing you to bulk-import products, suppliers, categories, and more without clicking through the UI one record at a time.

Overview

Instead of manually uploading import files through Tools > Imports in Luminous, you can hand Claude a spreadsheet and describe what you want done. Claude will read your file, ask clarifying questions (cost fields, categories, suppliers, etc.), submit the import job through the Luminous API, monitor it for completion, and automatically download and diagnose any error reports — all in one conversation.

Use this when you have large imports (e.g., 100–400+ products at once) and want to save time, reduce token usage from one-by-one imports, and get intelligent error handling without leaving the chat.

How It Works (Step-by-Step)

  1. Start a conversation with Claude — in Claude.ai or wherever you use it with the Luminous MCP connected.
  2. Tell Claude what you want to import. Example: "I want to do a bunch of imports in Luminous — here's my product file."
  3. Attach your file — upload the spreadsheet (CSV or Excel) directly into the chat.
  4. Answer Claude's clarifying questions. Claude will read your file and ask about things like:
    • Which cost field to use (if multiple are present)
    • How to handle categories that don't exist yet
    • Whether to create missing suppliers automatically
    • Any additional fields you want mapped
  5. Claude submits the import job. It calls the Luminous API and kicks off the job, returning a Job ID (e.g., Job #80).
  6. Verify in Luminous. Go to Tools > Imports > Import History to see the job listed, just like any manually submitted import.
  7. Claude monitors for errors. If the job fails or has row-level errors, Claude automatically downloads the error report — the same report you'd download manually from Import History.
  8. Claude diagnoses and resolves. Claude reads the error report, explains what went wrong, and proposes a fix. You can then re-run the corrected import in the same conversation.

Key Features / Options

Automatic error report download — Claude doesn't just tell you there's an error; it downloads the report, parses it, and tells you specifically which rows failed and why.

Intelligent field mapping — If your file has ambiguous or multiple cost fields, Claude asks which one to use rather than guessing wrong.

Auto-discovery of missing records — Claude can identify categories and suppliers referenced in your file that don't yet exist in Luminous, and handle their creation as part of the import flow.

Full import type support — This works with any import type available in Luminous: products/master SKUs, variant SKUs, suppliers, and more.

Job ID tracking — Every import returns a Job ID you can reference in Luminous's Import History for audit or manual verification.

Tips & Best Practices

Clean your file before uploading. Claude can handle some inconsistencies, but a well-structured spreadsheet with clear column headers will produce faster, cleaner results.

Let Claude ask its questions. Don't skip past the clarifying prompts — answering them correctly (especially around cost fields and categories) prevents avoidable import errors.

Use the Job ID. After Claude submits an import, copy the Job ID and verify in Tools > Imports > Import History. This is your audit trail.

For very large catalogs (1,000+ SKUs), consider the Unified Importer first. The Luminous Unified Importer supports up to 15,000 line items with a structured template and is optimized for that scale. Claude's AI import is best for conversational, flexible imports where you want intelligent field mapping and error handling.

Don't re-run the same import twice without checking. Claude will warn you if a supplier or category already exists, but duplicate SKU submissions can cause errors. Review Claude's summary before confirming.

FAQs & Troubleshooting

Q: What types of imports does Claude support? Any import type available in Luminous's UI — products/master SKUs, variant SKUs, suppliers, categories, BOMs, and more.

Q: Do I need a specific file format? No. Claude reads your file, maps the columns, and asks clarifying questions to handle variations. That said, the cleaner your file, the smoother the process.

Q: What happens if the job has errors? Claude automatically downloads the error report from Luminous, reads it, and tells you which rows failed and why — then suggests a fix so you can re-run in the same session.

Q: Will Claude create missing categories or suppliers automatically? It will ask you first. Claude identifies records that don't yet exist in Luminous and confirms with you before creating them.

Q: Can I still use the manual import UI? Yes — Tools > Imports in Luminous works exactly as before. This is an additional option, not a replacement.

Q: I copied the encrypted API token instead of the unencrypted one — what do I do? Delete the existing API key in Luminous under Settings > API Tokens, generate a new one, and copy the unencrypted token. Paste it into Claude when prompted. (This is the same issue that affects the Unified Importer — see the Unified Importer article for reference.)

Q: Why is this better than asking Claude to import records one by one? One-by-one imports are extremely slow, consume significantly more API tokens, and risk partial failures mid-batch with no clean recovery path. The bulk import approach submits a single job, tracks it as a unit, and handles errors holistically.

Additional Help

If you run into issues not covered here, contact the Luminous support team through the Help Center chat or email your account manager. When reaching out, include your Job ID from the import attempt so the team can look up the job directly.