RunLocalTools

Review Text Analyzer — Find Common Complaints & Praise in Bulk Reviews

Paste a batch of customer reviews and get instant word frequency, complaint/praise keyword signals, and a per-review breakdown — entirely in your browser.

What it does

Paste a batch of reviews — your own product’s, or a competitor’s you’re researching — and this tool breaks them down instantly: which words and phrases come up most often, which reviews mention common complaint or praise signals, and a per-review table you can filter by keyword. It’s built for spotting patterns across dozens or hundreds of reviews at once, faster than reading them one by one.

How it works

The tool splits your pasted text into individual reviews (by blank lines if present, otherwise one per line), and strips off star ratings or numeric scores like ★★★☆☆ or 4/5 so they don’t pollute the word counts. It then tokenizes the remaining text, filters out common English filler words (stopwords), and counts what’s left — both single words and short phrases. Separately, it checks each review against a bundled list of common complaint words (like “broke,” “refund,” “rude”) and praise words (like “great,” “recommend,” “durable”), using whole-word matching so a word like “fault” inside “cannot fault the build quality” doesn’t get miscounted just because it shares letters with something else. You can edit both lists to fit your own product category.

How to use it

  1. Paste your reviews into the text box — one per line, or separated by a blank line if each review spans multiple lines.
  2. Check the stats bar and the top-keyword list that appears automatically.
  3. Click any keyword to filter the review table down to just the reviews that mention it.
  4. Edit the negative/positive word lists in the sidebar if your product has its own common complaint terms.
  5. Export the word counts or the full per-review breakdown as a CSV, or copy a quick text summary.

How it’s different from AI sentiment analysis

This is deliberately simple, transparent keyword matching — not a machine-learning sentiment model. It counts words from a list you can see and edit, so you always know exactly why a review got flagged. Treat the negative/positive counts as a quick signal to point you toward reviews worth reading closely, not a final verdict.

Privacy

Everything runs in your browser. The reviews you paste are never uploaded anywhere — there’s no server involved in the analysis at all.

FAQ

Does this work with reviews copied straight from Amazon, Google, or the App Store? Yes — paste the raw text as you copied it. The tool automatically strips star-rating symbols and simple numeric ratings like “4/5” from the start of each review before analyzing it.

Can I add my own complaint or praise keywords? Yes — both word lists are editable text boxes in the sidebar. Add or remove comma-separated words to match the language your customers actually use.

Will it flag a word that’s part of a longer word, like “fault” inside “default”? No — matching is done on whole tokenized words, not substrings, so “fault” only counts when it appears as its own word.

Can I export the results? Yes — download the top word counts or the full per-review breakdown as a CSV, or copy a short plain-text summary to paste elsewhere.

Is there a limit to how many reviews I can paste? There’s no hard limit, but very large pastes (thousands of reviews) will take a moment longer to process since everything runs locally in your browser rather than on a server.

If you’re cleaning up a messy CSV export of reviews before pasting them in, the CSV cleaner can trim whitespace and dedupe rows first, and the word/character counter is handy for checking the length of an individual review you’re drafting a response to.