SENTIMENT ANALYSIS

Paste the reviews. See what they're really saying.

Customer sentiment scored, themed, and summarized. From noise to signal.

REVIEW INPUT

HOW IT WORKS

Under the hood

  1. 1
    Per-review scoring Each review is scored individually on a sentiment scale from -1.0 (very negative) to +1.0 (very positive).
  2. 2
    Theme extraction Topics like food quality, wait time, staff, value, and atmosphere are identified and grouped across reviews.
  3. 3
    Aggregate analysis Overall sentiment score, NPS estimate, and trend direction are calculated from the full review batch.
  4. 4
    Insight generation Claude surfaces the top praise, top complaint, and highest-priority action item from the data.
  5. 5
    Visualization render Results are rendered as a theme breakdown, sentiment bar, and review-by-review detail panel.

BUILT WITH

Claude API (NLP + sentiment)D3.js (visualization)Theme extraction pipelineCloudflare Workers AI
You just turned raw reviews into scored themes, trends, and actionable insights. In production, this monitors customer sentiment across every channel in real time.
Try Meeting Intelligence to see sentiment extracted from live conversations.

More where that came from.

Back to all demos →