Analyze customer reviews in bulk
Drop in an export from your store, app marketplace or survey tool. Jev reads every review and labels its sentiment, the star rating it implies, what it is mostly about and whether the customer wants a refund. You see the split at a glance and can sort to the angriest reviews first.
What you upload
A CSV with one review per row, in a column such as review, text or comment.
What you get back
- Sentiment split: positive, mixed or negative
- Implied star rating, even when no rating was given
- Main topic: product quality, shipping, support or price
- Which reviews ask for a refund
How it works
- 1
Upload your file
A CSV with a header row, up to 1,000 rows and 2 MB. Or start with the sample data to see the result first.
- 2
Run it
You see the credit cost before anything is charged. Rows run in chunks with live progress, and failed rows are refunded.
- 3
Read and export
A summary chart, a sortable table and a CSV with every answer and its probability.
Each row is one evaluation: 1 credit for up to 8,000 characters and 5 questions, 3 credits above that. New accounts start with 200 free credits.
Run it on your data
Answers are model judgments with a probability attached. Check a sample of rows before acting on the full result.