TableMarkerUpload your table
Upload Excel or CSV files, with headers and delimiter as needed.
AI for Excel and CSV spreadsheets
row by row
AI for spreadsheets that runs your prompt across every row. Classify, enrich, extract and generate columns from Excel or CSV — no code, no sign-up.
How it works
Upload a table, write a prompt, and get results in the same format. Great for generation, classification, extraction, data work, and calculations.
1
2
3
4
Describe the task
Write a prompt. The model will run your task row by row using the table.
Pick an AI model
Before you run, choose the model you want for the job.
Get the result
Wait time depends on table size. When it’s done, download your processed table.
Many tasks, one tool
- Assign a product category from its description
- Classify orders by volume
- Detect the language of text in a row
- Normalize data to one format
- Extract ticket topic from chat
- Multiply values across columns
- Write sales copy from row data
- Extract keywords from text
- Pull company names from text
- Normalize date formats
- Check rows against criteria
- Generate a title from content
- Suggest clothing size from measurements
- Draft a reply to a marketplace review
- Parse region from column data
- Add weekly forecast columns
- Build a lightweight user profile
- Prioritize backlog tasks
- Generate replies to support tickets
- Price orders from a rate card
Data security
We don’t store your tables on our servers. Processing happens in memory; you get the result — source files aren’t saved.
Please don’t upload personal data, trade secrets, classified information, or other sensitive content. The service isn’t meant for that, and we aren’t responsible for such data.
FAQ
Excel (.xlsx, .xls) and UTF-8 CSV. For best results, keep columns consistent, avoid merged cells, and don’t embed media in cells.
We’re still improving this. We can’t guarantee correctness yet because the model sees rows one at a time, not the full dataset at once.
Limits depend on your plan. On the free tier, row count is capped (e.g. up to 20). Paid plans offer higher limits.
We don’t store your tables: they’re processed in memory and not saved on our servers after the job. Data is still sent to external AI providers for processing, so avoid personal or confidential content.
Run processing multiple times — each time use the latest output as the new input — once per extra column you need.
Upload an .xlsx or .xls file, mark headers, and describe the task in a prompt — for example, “tag this product by category”. The model reads each row and writes the result into a new column. Download the processed file back as Excel when it’s done.
Yes. The free tier lets you upload a CSV and process up to 20 rows without registration. Prompts can extract entities, normalize formats or split values into separate columns.
ChatGPT is conversational and struggles with large tables. Excel can compute but doesn’t understand text. TableMarker is AI for spreadsheets: it applies the same prompt to every row, keeps the table structure intact, and needs no formulas, macros or code.
List the categories and rules in your prompt. For each row the model reads the chosen columns and writes the picked label into a new column. Great for ticket tagging, lead scoring, product categorization and review themes.
Export reviews into a table, set tone and template in a prompt — TableMarker drafts a reply for each row in a new column, in bulk and in one consistent voice. No copy-pasting into chat.
Many tasks, one tool
Try processing your table for free, no sign-up required.