Sentiment label + confidence
Each text is classified as positive, negative or neutral, with a confidence score from 0 to 1.
CatalogLatency Typically 0.5–2 seconds Async false Rate Limit Per API key Max Input A few thousand characters per text
cn8/text-sentiment
Send a piece of text and get its sentiment back: a label (positive, negative or
Learn more about Sentiment Analysis Text NLPsyncInstant previewtextanalysisnlp
Try it
Texts to analyzeOne text per line — each line gets its own verdict.
Result
Results show up here
The sample input is ready — just hit Run.
Overview
Send a piece of text and get its sentiment back: a label (positive, negative or neutral), a confidence score, a continuous polarity from -1 to +1, a 1–5 star rating, and the detected language. Responses are synchronous — one request, one result. Ideal for scoring reviews, comments, support messages and survey answers.
Key capabilities
Polarity score
A continuous value from -1 (very negative) to +1 (very positive) for finer-grained ranking.
1–5 star rating
A familiar star rating ("1 stars" … "5 stars") derived from the polarity — handy for review dashboards.
Automatic language detection
The detected language is returned with every result, so you don't need a separate detection step.
When to use it
Review & rating analysis
Score customer reviews and map them to stars for dashboards and alerts.
Social listening
Track positive and negative trends across mentions, comments and posts.
Support triage
Surface negative messages in support queues so they're answered first.
Input & output
input
A single text to analyze (sent as a one-element texts array).
JSON body
output
One result with label, confidence score, polarity, star rating, is_positive and detected language.
JSON
Guides & tips
How it works
- Your text is classified as positive, negative or neutral by a sentiment model.
- Polarity is a continuous reading from -1 to +1; the star rating maps that polarity to a familiar 1–5 scale.
- The response also reports the language it detected, so you can branch on it without a separate call.
Tips
- Use the field name
texts(an array), and send one text per request. - To analyze many texts, call the endpoint once per text and collect the results.
- For finer-grained emotions (joy, anger, …) see Emotion Detection; for your own custom labels see Zero-Shot Classification.
Specs
Schema
Request body
textsrequiredarray<string>The text to analyze, as a one-element array. Send one text per request.
Response
results[].labelstringSentiment label: positive, negative or neutral.
results[].scorenumberConfidence of the label, from 0 to 1.
results[].starsstring1–5 star rating as a string, e.g. "5 stars".
results[].is_positivebooleanTrue when the label is positive — a convenience flag.
results[].polaritynumberContinuous polarity from -1 (very negative) to +1 (very positive).
results[].languagestringDetected language of the text (ISO 639-1).
Pricing
Billed per request.
| Service | Unit | Price |
|---|---|---|
| Sentiment Analysis | item | $0.002/request |