Sentiment Analysis API

About the Sentiment Analysis API

A sentiment analysis API and text analysis API for reviews, tickets and social posts. Send any text and get back a sentiment score, a five-step label from very negative to very positive, and isPositive and isNegative flags you can route on without choosing thresholds yourself.

Every plan, including the free one, scores text against the AFINN word list. Each known word carries a value from -5 to +5, the values are added up, and a comparative score divides by the number of words so a one-line tweet and a long review can be compared fairly. It is fast and fully explainable, which suits dashboards, alerts and bulk tagging of customer feedback.

Paid plans add a transformer model that reads the text sentence by sentence. It returns its own label and a -1 to 1 score, flags mixed text where one sentence is clearly positive and another clearly negative, and lists the words that moved the score. That is where social media sentiment analysis and support triage gain the most: "great app, but billing is broken" comes back as mixed instead of neutral.

How it works

Sentiment Analysis runs two independent scorers over the text you POST. The first is lexicon-based: the text is tokenised, each token is looked up in the AFINN list of integer valences from -5 to +5, and the matches are summed into sentiment. comparative divides that sum by the token count. sentimentText maps the sum to a label with fixed cut-offs: 4 and above is very positive, 2 to 3 positive, -1 to 1 neutral, -2 negative, and -3 or below very negative. isPositive and isNegative are only the sign of the sum, so a score of 1 returns isPositive: true while the label still reads neutral.

On paid plans the text is also split into sentences at terminal punctuation and line breaks, and the first 50 sentences go to a transformer classifier trained on social media text. Each sentence gets positive, neutral and negative probabilities. modelScore is the mean positive probability minus the mean negative probability, rounded to two decimals, and modelSentiment is the label with the highest mean. isMixed is true when at least one sentence is labelled positive and another negative, each with confidence of 0.6 or more; modelSentiment then reads mixed.

normalizedScore is the AFINN sum divided by 10 and clamped to -1..1, so any text scoring 10 or more in either direction saturates. magnitude is the sum of absolute valences and grows with length: use it to tell strongly worded text from flat text, not to compare documents of different sizes. The free plan skips the model call entirely, so the model fields are null there.

What people use it for

Support Ticket Escalation

Route angry customer messages to senior support reps immediately when the sentiment score drops to very negative.

Review Stream Filtering

Product managers categorize incoming app store feedback into positive praise and negative complaints using the five-tier sentiment label.

Social Media Sentiment Analysis

When tracking brand tags on social networks, marketing teams flag sudden spikes in negative valence scores to address PR issues early.

Survey Response Sorting

Sort open-ended employee survey responses automatically by comparative sentiment to isolate recurring workplace grievances without manual reading.

Ways to call it

One endpoint, many ways in — REST with JSON, XML, YAML and CSV, plus GraphQL and an MCP interface for AI agents.

JSON

Default REST response

XML

Markup format

YAML

Human-readable

CSV

Tabular export

Beta

GraphQL

Query language

New

MCP

For AI agents

Request parameters — POST /v1/sentimentanalysis

Sent as JSON in the request body. Premium parameters are accepted on every plan but only take effect on plans that include them. Anything not listed here is dropped rather than passed through.

Required

1
2
3
4

curl -X POST "https://api.apiverve.com/v1/sentimentanalysis" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text":"I'\''m so excited that tomorrow is going to be sunny! Can'\''t wait!"}'

Authentication

Send your key in the x-api-key header. That is the only auth step — no token exchange, and no per-endpoint scope to configure.

Response — POST /v1/sentimentanalysis

Every APIVerve endpoint returns the same three top-level keys, so one response handler covers your whole integration: status, error and data. Only data changes shape.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

{
  "status": "ok",
  "error": null,
  "data": {
    "comparative": 0.25,
    "sentimentText": "positive",
    "sentiment": 3,
    "isPositive": true,
    "isNegative": false,
    "normalizedScore": 0.3,
    "magnitude": 3,
    "positiveWords": [
      "excited"
    ],
    "negativeWords": [],
    "modelSentiment": "positive",
    "modelScore": 0.98,
    "isMixed": false,
    "sentences": [
      {
        "text": "I'm so excited that tomorrow is going to be sunny!",
        "label": "positive",
        "confidence": 0.992,
        "scores": {
          "positive": 0.992,
          "neutral": 0.006,
          "negative": 0.002
        }
      },
      {
        "text": "Can't wait!",
        "label": "positive",
        "confidence": 0.973,
        "scores": {
          "positive": 0.973,
          "neutral": 0.024,
          "negative": 0.003
        }
      }
    ]
  }
}

Response fields

Paths are relative to data. Premium fields are absent rather than zeroed on plans that do not include them, so check for presence instead of comparing to 0.

Response headers

Every response carries your balance, so your own code always knows where it stands without polling anything. All four are exposed to browsers through CORS.

Errors

Read the HTTP status first, then error for the specific reason. The body names the parameter that has to change.

What a call costs — 10 credits per call

Credits are shared across every APIVerve API — one balance, one key, one bill. Failed requests never count against it, so a rejected input costs nothing.

Calls are what the credits buy at this API's rate — spend them here, on any of the other APIs, or across both. Nothing is reserved per endpoint.

Zero-code embed — one snippet, no backend

Drop an interactive Sentiment Analysis form onto any page. The widget calls the API for you, so no key ever appears in your markup and there is nothing to deploy.

Changelog — Sentiment Analysis API

Every change to this endpoint, newest first. Breaking changes ship as a new version and the previous one keeps serving — response fields are added, never removed or retyped in place, so an integration written against v1 keeps working.

Currently v1