Turn product photos into clean catalog data with an AI driven flow connected to Airtable. Ideal for inventory audits, marketplace listings, and field surveys, it finds product names and key attributes, then writes results back to your table.

The run starts by reading Airtable rows that have an image and are not yet processed. An AI vision step describes the photo. An agent powered by a chat model then researches the web. It can run a Google reverse image search through SerpAPI and scrape pages with Firecrawl, returning short markdown summaries. A router chooses the right tool, handles success and errors, and sends a compact response back to the agent. A structured output parser forces clean fields. Finally, the enriched data is saved to Airtable so your records stay current.

Setup needs your Airtable base and table IDs plus API keys for OpenAI, SerpAPI, and Firecrawl. Expect big time savings as manual research drops from many minutes per item to a few minutes. This is strong for teams that collect photos in the field and need quick, accurate product data. After connecting credentials and mapping fields, test a few images, review the updates, and then scale to larger batches.

• Filters Airtable to only process rows with images and not yet marked as complete.
• Uses an AI vision model to describe each product photo in clear language.
• Runs an agent with a chat model and a structured output parser for clean fields.
• Calls a custom reverse image search tool through SerpAPI to find matching pages.
• Scrapes product pages with Firecrawl and returns markdown for fast parsing.
• Routes tool requests with an execute trigger, switch logic, and fallback handling.
• Writes enriched names and attributes back to Airtable to keep records current.

• Reduce manual product research from 30 minutes to 3 minutes per item
• Streamline catalog enrichment by up to 80 percent using AI and web tools
• Improve data accuracy by checking details against live web sources
• Handle up to 10 times more products with the same team size
• Connect Airtable with OpenAI, SerpAPI, and Firecrawl in one flow

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Airtable, OpenAI, SerpAPI and Firecrawl. See the Tools Required section above for links to create accounts with these services.
3. In Airtable, create or open a base and table with at least these fields: Image as an Attachment field, AI_status as a Checkbox or Boolean field, and any product fields you want to update. Note the Base ID and Table ID.
4. In the n8n credentials manager, create an Airtable Personal Access Token credential. Generate a token in your Airtable account, paste it into n8n, and save it with a clear name.
5. Create an OpenAI credential in n8n. In your OpenAI account, create an API key, paste it into n8n, and save. Assign this credential to both the chat model and the vision analysis nodes.
6. Create a SerpAPI credential in n8n. In your SerpAPI dashboard, copy your API key and add it to the SerpAPI credential. Save it and link it to the reverse image HTTP Request node.
7. Create a Firecrawl HTTP Header credential in n8n. Set Header Name to Authorization and Value to Bearer YOUR_FIRECRAWL_KEY. Save and select it in the Firecrawl Scrape API node.
8. Open the Airtable nodes. Set the Base ID and Table ID to your own. Keep the filter formula so only rows with an image and AI_status false are fetched.
9. Open the AI nodes. In the chat model, select model gpt 4o and temperature 0. In the vision step, use the same OpenAI credential and confirm the image input is mapped from Airtable.
10. Check the tool routing. Leave the route values and tool names as provided so the Actions Router can send requests to SerpAPI or Firecrawl correctly.
11. Run a test with a few rows. Click Test workflow, watch executions, and confirm the reverse image or scrape response appears in the response field.
12. Verify Airtable updates. Confirm enriched fields are written to the correct records and AI_status is updated if configured. If no updates appear, recheck Base ID, Table ID, and field names.
13. Troubleshoot common issues: if Firecrawl fails, check your key and URL reachability; if SerpAPI returns no results, ensure the image URL is public; if the agent stops with a fallback message, confirm route names match the router options.