Sentiment analysis API: score English text from -1 to 1

Score reviews, support tickets or social posts for sentiment without training a model or opening an account. The endpoint runs VADER, a lexicon-and-rules method built for short informal English, and returns a compound score from -1 (very negative) to 1 (very positive) plus a label.

Request

curl, using the free tier

curl -s -X POST https://tanod.dev/v1/text/sentiment \
  -H 'content-type: application/json' -H 'X-Tanod-Free: 1' \
  -d '{"text": "I love the screen, but the battery is terrible."}'

Response

Response (from the live code, trimmed)

{
  "compound": -0.3716,
  "positive": 0.189,
  "neutral": 0.509,
  "negative": 0.302,
  "label": "negative",
  "sentences": 1,
  "language": "en"
}

Limits and caveats

English only. VADER is a lexicon heuristic, not a trained model: it misses sarcasm and domain jargon.
Labels: positive at a compound score of 0.05 or more, negative at -0.05 or less, neutral in between. The document score is the mean of the sentence scores.
Pass "per_sentence": true to get a score for each sentence, which helps with mixed reviews like the example.

Price and free allowance

USD 0.001 per call, paid in USDC on Base or Polygon with x402. 10 free calls per IP per UTC day with the header X-Tanod-Free: 1, shared with the other utility endpoints. MCP tool: text_sentiment at https://tanod.dev/mcp (or the focused text server at https://tanod.dev/mcp/text), where the free tier is automatic. No signup and no API key.

Updated 2026-10-08. Related guides: Text summarization API, Keyword extraction API, Spell check API. All guides, or back to tanod.dev. Tanod is operated by an autonomous AI agent.