Enhance scraped data with AI
Simplescraper can enhance the data you scrape using AI: clean it up, translate it, or pull out details buried in longer text, such as an email address in a page's text or the salary range in a job description. Each answer goes into a new column beside your data, one for every row, and your scraped data is never changed.
AI Enhance can:
- Extract one detail, such as an email or a salary range
- Summarize long text into a line or two
- Answer yes or no to your question, such as "Does this job involve travel?"
- Sort each row into one of your categories
- Follow your own instruction, such as translating a description or cleaning up a title
You can add AI columns to a saved recipe in the dashboard, or add AI fields in the Chrome extension while you scrape.
Add AI columns in the dashboard
This video adds a Categorize column that sorts YouTube Shorts into B2B and B2C:
In the Simplescraper dashboard, open a recipe's 'Results' tab and click 'AI Enhance' above the table.
Pick a starting point from the list, or click 'New AI column' if the recipe already has one.

Type a name such as
is_remotein 'New AI column name', fill in the column type's fields and tick the properties the AI should read under 'Properties to include'.
Click 'Preview on 3 rows' to check the answers.
Click 'Save', or 'Save and run' to fill the column for every row now.
To fill the column every time the recipe runs, including scheduled runs, set 'When it runs' to 'After each scrape' before you save.
What each type of AI column does
The examples below use a recipe that scrapes job listings.
A Summarize column writes a short summary of the properties you choose.
- Reads
job_description, a 600-word posting, with 'Length' set to 'Two bullets' - Writes 'Builds data pipelines for the billing team' and 'Owns reporting for finance and sales'
An Extract details column finds one detail in each row and writes it to the column.
- Reads
job_descriptionand asks for 'The email address to apply to' - Writes 'careers@example.com'
A Yes or no column answers one question about each row.
- Reads
job_titleandlocationsand asks 'Does this job allow remote work?' - Writes 'Yes' for a job in 'Remote - US' and 'No' for a job in 'Austin, TX'
A Categorize column puts each row in one of the categories you list and never invents another.
- Reads
departmentwith the categories 'Engineering', 'Research' and 'Sales' - Writes 'Engineering' for 'Applied AI Engineering' and 'Sales' for 'Technical Success'
Yes or no and Categorize columns fill every row with the answer the AI finds most likely. A Categorize column also offers 'Add an 'Other' category for rows that fit none', ticked for a new column, so a row about something outside your list is marked 'Other' instead of being put in the nearest category.
A Custom instruction column follows your own instruction for every row.
- Reads
job_titleand follows 'Spell out abbreviations in the title' - Writes 'Senior Software Engineer (Payments)' for 'Sr. Software Engineer (Payments)'
Credits
AI Enhance uses 1 credit per 800 words, counting the data it reads plus the answer it writes. The panel shows an estimate before you preview or run.
Add AI fields in the Chrome extension
In the extension, you add AI fields while you build a recipe, and they fill in for every row each time the recipe runs.
- On the 'Confirm fields' step, click 'Add an AI field'.
- Give the field a name and choose what it does:
- Extract details: pulls one value out of the text, such as the lowest base pay.
- Categorize: picks the best match from a short list you write, such as the degree a job requires.
- Custom instruction: follows your own instruction, such as listing the key skills a job asks for.
- Yes or no: answers one question, such as whether a job involves travel.
- Describe what you want: the value to extract, the categories, the instruction or the question.
- Under 'Fields to read', tick the fields that hold the data.
- Save the field.
This video adds four AI fields on Apple's job pages (the lowest salary, the degree, the key skills and whether the job involves travel), then runs the recipe on ten job URLs:
AI columns in your integrations
- Google Sheets: the new column is added to the right of your existing columns.
- Email: AI columns are included unless you switch one off in the email panel.
- Airtable: switch the column on in the Airtable panel and add a field with the same name to your table. Without that field, the column is skipped, every other field still arrives and the panel says which field was skipped.
What AI Enhance cannot do
AI Enhance reads only the data already in each row, so it can't click, open links, move to the next page or search the web.
Run the recipe once first
In the dashboard, AI Enhance works on data that is already scraped, so the recipe needs at least one run. It doesn't need a full run: click 'Test' on the recipe page to scrape a single page, and AI Enhance can use those rows.
FAQ
Does AI Enhance change my scraped data?
No. Each AI column writes only to its own column, and deleting a column keeps the values it already wrote.
What happens when a row doesn't contain the answer?
It depends on the column type:
- Summarize and Custom instruction: the cell stays empty.
- Extract details: the cell follows the 'If the detail is not there' setting, 'Leave empty' or "Write 'Not stated'".
- Yes or no: the column always answers, writing 'No' or 'false' depending on 'Answer as'.
- Categorize: every row gets the category it fits best. With 'Add an 'Other' category for rows that fit none' ticked, a row that fits none of yours gets 'Other'.
How many AI columns can a recipe have?
Up to 10.
What happens to my older AI Enhance setup?
It keeps working as before. It appears first in the panel's list, marked 'Legacy'.
More Guides
- Credits - how credits are counted across Simplescraper
- Scheduling your webscraping - pair a schedule with columns that run after each scrape
- Sub-page (deep) scraping - bring linked pages into your rows so AI Enhance can read them
For questions or help, reach out via chat.