Summarize reviews
A 4.1-star average can hide customers who love the product and hate the shipping. Feed in the reviews and get the recurring praise and complaints by theme, with the phrases customers repeat, not one loud outlier. You pay per use, with no subscription.
Run it online
Run this on our servers with your account. Free tools run in your browser; this one bills your KIT balance per the price above.
Nobody reads five hundred reviews start to finish, so most of what customers actually say gets skimmed at three stars and ignored.
What a star average cannot tell you
Nobody reads five hundred reviews from start to finish, so most of what customers say gets skimmed and ignored. A 4.1-star rating says almost nothing about whether customers love the product and dislike the shipping, or love the shipping and are lukewarm about the product: two very different businesses can share the same number. The difference is written in plain language inside the reviews, but only someone who reads them all will find it. This tool reads the whole batch, groups the comments by theme and separates what is working from what is genuinely broken.
Patterns, not one loud voice
The summary favors the praise, complaints and specific details that show up again and again across many reviews, rather than a single angry or glowing outlier. That makes it useful for product teams deciding what to fix next, for sellers trying to understand a sudden drop in their rating, and for brands with several product lines that want to see which one has a genuinely different set of recurring issues. A weekly run over new reviews gives a running digest instead of spot checks that may not be representative. The reviews can come from a marketplace export, a review site or your own feedback form.
The star average lies by omission
A 4.1-star rating tells you almost nothing about whether customers love the product and hate the shipping, or love the shipping and are lukewarm on the product itself — two very different businesses that can share the same number. Reviews carry that distinction in plain language, but only if someone actually reads them, which at real volume nobody consistently does.
Who reaches for this
Product teams deciding what to fix next without guessing from a handful of recent reviews they happened to skim. Marketing teams looking for the specific language customers use, which is almost always better copy than anything written in-house. Sellers on a marketplace trying to understand a sudden rating drop across a batch of new reviews without reading every one manually.
What you can do with it
Weekly product digests
A product team gets a running summary of new reviews each week instead of spot-checking a handful and hoping they're representative.
Understanding a rating drop
A seller whose average rating dropped this month feeds in recent reviews to see exactly which recurring complaint is driving it.
Mining customer language for copy
A marketing team pulls recurring phrases from summarized reviews to use the exact words customers already use to describe the product's value.
Comparing feedback across product lines
A brand with several product lines summarizes reviews for each separately to see which one has a genuinely different set of recurring issues.
FAQ
Is there a free trial?
There's no free tier — free tiers get abused and slow everyone down. No subscription, no tokens, no invented credits, and a failed task is never charged.
Does it separate positive from negative themes?
Yes, the summary groups recurring praise and recurring complaints separately so you can see both sides without them blending together.
Am I charged if the summarization fails?
No. Failed tasks retry automatically up to three times and are never charged; you only pay for summaries actually delivered.
Can it explain why my rating dropped?
Yes. Feed in the recent reviews from the period of the drop and the summary shows which recurring complaint is driving it, instead of leaving you to guess from the few reviews you happened to read.
Can I use the summary for marketing copy?
Yes. It surfaces the phrases that recur across many reviews, so a marketing team can reuse the exact words customers already use to describe what they value, which is almost always better than copy written in-house.
Related tools
For developers — API access
Everything on this page is available programmatically. This section is for teams who want to wire it into their own systems; everyone else can just use the tool above.
API endpoint
Prefer to automate it? One authenticated POST creates the task; the result comes back by webhook or a signed link. The same capability also runs here on the web, by email and from Telegram — and soon from our app too.
Call it from your stack
curl -X POST https://api.kit.forhosting.com/text/review-summary \
-H "Authorization: Bearer $KIT_KEY" \
-H "Content-Type: application/json" \
-d '{"text":"El informe trimestral cerró con 1,2 M€ de ingresos."}'const res = await fetch("https://api.kit.forhosting.com/text/review-summary", {
method: "POST",
headers: {
"Authorization": `Bearer ${process.env.KIT_KEY}`,
"Content-Type": "application/json"
},
body: JSON.stringify({
"text": "El informe trimestral cerró con 1,2 M€ de ingresos."
})
});
const { task_id } = await res.json();import os, requests
res = requests.post(
"https://api.kit.forhosting.com/text/review-summary",
headers={"Authorization": f"Bearer {os.environ['KIT_KEY']}"},
json={
"text": "El informe trimestral cerró con 1,2 M€ de ingresos."
},
)
task_id = res.json()["task_id"]<?php
$res = file_get_contents("https://api.kit.forhosting.com/text/review-summary", false, stream_context_create([
"http" => [
"method" => "POST",
"header" => "Authorization: Bearer " . getenv("KIT_KEY") . "\r\nContent-Type: application/json",
"content" => '{"text":"El informe trimestral cerró con 1,2 M€ de ingresos."}',
],
]));
$task = json_decode($res, true);body := bytes.NewBufferString(`{"text":"El informe trimestral cerró con 1,2 M€ de ingresos."}`)
req, _ := http.NewRequest("POST", "https://api.kit.forhosting.com/text/review-summary", body)
req.Header.Set("Authorization", "Bearer "+os.Getenv("KIT_KEY"))
req.Header.Set("Content-Type", "application/json")
res, _ := http.DefaultClient.Do(req)Example request
{
"text": "El informe trimestral cerró con 1,2 M€ de ingresos."
}Example response
{
"task_id": "tsk_a1b2c3d4e5f6a1b2c3d4e5f6",
"type": "text.review_summary",
"status": "queued",
"_links": {
"result": "/tasks/tsk_…/result"
}
}The API is asynchronous: the call returns a task_id immediately and the result arrives by webhook. Polling is capped at 1 req/s per task.
Pricing
Published price — no tokens, no invented credits. A failed task is never charged.
Limits
max_tokens | 20000 |
Errors
| HTTP | Code | Meaning |
|---|---|---|
401 | unauthorized | Missing or invalid API key. |
402 | insufficient_balance | Your balance doesn't cover the task price. |
404 | unknown_type | That task type doesn't exist. |
429 | rate_limited | Too many requests. Use the webhook instead of polling. |
422 | task_failed | The task failed after 3 retries. You are never charged for it. |
Technical notes
This endpoint reads the full batch and returns a genuine summary — recurring praise, recurring complaints, and the specific details that show up again and again — instead of a star average that flattens all of it.
What goes in, what comes back
The output reads like a summary a careful person would write after actually reading all of them, because that's functionally what happened. You POST a batch of review text — pulled from a marketplace export, a review platform or your own product feedback form — and the task groups them by theme, surfaces the phrases and specifics that recur across many reviews rather than one loud outlier, and separates what's working from what's genuinely broken.
Reviews as a genre have their own history
Consumer reviews as a mass phenomenon are barely three decades old, born with early online marketplaces that let strangers rate strangers' products at scale for the first time — before that, word of mouth was slow and local. That scale is exactly what makes reviews valuable and exactly what makes them unreadable past a certain volume; this endpoint exists at the point where the value and the unreadability collide.
Feeding an automated feedback loop
Because the summary returns as structured data through a webhook, it can populate a weekly product digest, trigger an alert when a new theme of complaints emerges, or feed directly into our feedback analysis endpoint for deeper theme and sentiment breakdowns, without anyone manually exporting a spreadsheet first.
Technical questions
How does the review summarizer API work?
You POST a batch of review text to /text/review-summary, the task runs asynchronously, and you receive a thematic summary of recurring praise and complaints via webhook or signed link.
How many reviews can I summarize at once?
Reviews are billed per request plus per batch, so you can submit a large batch at once and the price scales with the number of reviews included.
What formats can I submit reviews in?
Plain text works directly; exports from marketplaces or review platforms just need the review text extracted into the request.
How is this different from just reading the star rating?
A star average collapses everything into one number; this endpoint returns the actual recurring reasons behind that number, in the customers' own words.
Can I combine it with sentiment or theme analysis?
Yes, review summaries pair naturally with our feedback analysis endpoint for a deeper breakdown of sentiment and priority per theme.