Turn product photos in Airtable into clean, searchable data. Teams that manage catalogs, inventory surveys, or marketplace listings can use this to identify items, fetch details from the web, and update records without hours of manual research.

The flow starts with a manual run and pulls rows from Airtable that have an image and are not yet processed. An AI vision model describes the photo. An agent then uses two tools. The first tool performs a reverse image search through SerpAPI to find matching pages. The second tool scrapes those pages with Firecrawl and returns readable text. A structured parser shapes the results into fields, and the workflow writes the enriched data back to Airtable. Error checks handle scrape failures and provide a fallback response. The design uses a tool router so the agent can choose the right action.

Setup needs your Airtable base and table, plus API keys for OpenAI, SerpAPI, and Firecrawl. Expect faster listing prep and fewer data entry mistakes. Use it for store audits, product intake from field surveys, or to prepare marketplace listings from photos. After testing on a few rows, expand to larger batches when quality meets your standard.

• Airtable search filters rows with images and a false status so only new items are processed
• Vision analysis uses an OpenAI chat model with low randomness for stable descriptions
• Agent chooses tools through a router so it can reverse image search or scrape pages when needed
• Reverse image search calls SerpAPI with the image url to find matching web pages
• Web scraping uses Firecrawl and returns page content in markdown for easy parsing
• Success and error paths handle scrape results and provide a clean fallback message
• Structured output parser shapes AI results into consistent fields
• Writes enriched fields back to Airtable so the catalog stays updated

• Reduce manual product research from 30 minutes per item to about 2 minutes
• Automate up to 80 percent of repetitive enrichment work by routing tasks to the right tool
• Improve data quality by pulling details from multiple sources and using a structured parser
• Handle ten times more product photos with the same team size
• Connect Airtable, SerpAPI, OpenAI, 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 the n8n credentials manager, create an Airtable Personal Access Token credential. In Airtable, generate a token with read and write access to the base that holds your table.
4. Open the Airtable node that reads rows. Select your Airtable credential. Set the Base and Table fields. Keep the filter formula AND(Image!="", AI_status=FALSE()) or adjust to match your column names.
5. In the n8n credentials manager, create an OpenAI API Key credential. In your OpenAI account, create an API key and paste it. Select this credential in the AI model node.
6. In the n8n credentials manager, create a SerpAPI credential. In your SerpAPI dashboard, copy the API key and paste it. Select this credential in the reverse image HTTP Request node.
7. In the n8n credentials manager, create an HTTP Header Auth credential for Firecrawl. Add the Authorization header with your API key value. Ensure Content Type is application/json in the node headers.
8. Open the tool workflow nodes for reverse image and web scraper. Confirm the route fields match the router logic and that the workflow id reference is correct.
9. Open the Structured Output Parser node and review the expected fields. Make sure your Airtable table has matching columns for title, brand, url, description, and status.
10. Open the Airtable write back node. Map each parsed field to the correct Airtable column. Save the node.
11. Click Execute to run a test. Confirm the Scrape Success branch fires when Firecrawl returns data. If you see the fallback message, check that the image url is public and that all API keys are valid.
12. Validate a few updated rows in Airtable. If results look good, increase batch size by removing any row limits and run again. Add a schedule trigger later if you want regular runs.