Paste customer feedback, emails, social media posts, reviews, or survey responses and instantly understand the emotional tone.
Sentiment analysis measures whether a piece of writing is positive, neutral or negative. Businesses use it to understand customer reviews, survey responses, support emails and social media mentions at a glance.
Everything runs in your browser using open-source language tools. Your text and documents never leave your device, and the tool is completely free.
From text to context
Read the language behind a customer rating, not just the number.
Sentiment is not NPS. NPS depends on the share of promoters and detractors from 0–10 ratings; this tool only analyzes written feedback.
Analyze your textInspect what customers say about a product, delivery, or service.
No Amazon connection, scraping, or automatic review import is included. Mixed reviews can produce a neutral score even when they contain a serious complaint.
Analyze your textUse negative wording and frustration cues as a prompt for a closer read.
A calm or formal complaint can still be urgent. Scores do not determine severity, identify threats, or automatically escalate tickets.
Analyze your textSpot language cues in individual posts or comments you supply.
This is manual text analysis, not a live social feed. There is no platform connection, trend tracking, or automated monitoring.
Analyze your textUnderstand the labels
Tone describes expression; sentiment describes polarity. These rule-based cues can overlap. Explore the difference.
Work-oriented, courteous language about tasks or business.
please · timeline · project
Warm, appreciative, or welcoming wording.
thanks · glad · welcome
Excitement or strong praise; exclamations can reinforce a positive score.
amazing · excited · thrilled
Hostile or strongly disapproving wording; capitals and exclamations can reinforce a negative score.
furious · unacceptable · hate
Dissatisfaction, repeated problems, or waiting for a resolution.
waiting · disappointed · not working
The fallback when no other tone rule matches, not proof of an absence of emotion.
No matching tone cues
Conventional or official phrasing; longer text without certain contractions can also trigger this rule.
hereby · furthermore · sincerely
Informal, conversational language or common shorthand.
lol · gonna · btw
Strength comes from the absolute sentiment score: −70 and +70 are both Strong. It measures score magnitude, not emotion intensity or certainty.
Common questions
The open-source sentiment library assigns scores to recognized words and emoji and adds them together. This analyzer divides that total by the square root of the token count, then applies a soft scaling formula: round(100 × tanh(total ÷ (3 × √tokens))). The result ranges from −100 to +100. Scores above +5 are Positive, below −5 are Negative, and −5 through +5 are Neutral. This is a rule-based estimate, not a confidence percentage.
You can paste several responses together or upload a document containing multiple reviews, but you will receive one combined result for all the text. There is no batch upload, CSV import, or separate score per response. For individual results, analyze each response separately. Combined results can hide differences between happy and unhappy customers; the tool does not calculate an NPS score.
Customer feedback guideThe analyzer reads your files and calculates results locally in your browser. The analysis process does not upload your text or documents to a server, save them to an account, or send them to an AI API. Text stays in the current page's memory and is cleared when you reload or close it. The website still loads its page resources over the internet, so local analysis is not a promise that the entire website makes no network requests.
The analyzer labels sentiment Positive, Neutral, or Negative, assigns one of five strength levels, and checks for Professional, Friendly, Enthusiastic, Angry, Frustrated, Formal, and Casual language. It returns Neutral tone when no other tone rule matches and shows up to three matching tones. It does not independently classify emotions such as sadness, fear, or surprise. Tone labels are language cues, not a diagnosis of someone's feelings.
View all tones and strengthsSentiment polarity describes whether wording leans positive, neutral, or negative. Tone describes how it is expressed, such as formal, friendly, or frustrated. A politely written complaint can be Professional and Negative at the same time. Tone percentages reflect the relative rule matches, not the probability that a tone is correct, and the displayed top three may not add up to 100%.
You can upload a text-based PDF, a Word .docx document, or a plain .txt file. Older .doc files and spreadsheets are not supported. PDF extraction reads embedded text; it does not perform OCR, so a scanned or image-only PDF may contain no readable text. Copy text from the original document or use OCR separately before pasting it here.
This tool uses a word-scoring library and custom tone rules rather than a contextual AI model. Sarcasm, slang, specialist vocabulary, and mixed opinions can produce misleading results, while unfamiliar words may not affect the score at all. Positive and negative wording can cancel out. Read the original text and detected keywords alongside the score, especially before making decisions about customers or support escalations.
The default sentiment vocabulary and custom tone rules used here are English-focused. You can paste other languages, but this app does not detect the language or load language-specific dictionaries, so those results may be unreliable or Neutral. Use English text for the intended workflow; if you translate a response first, remember that translation can change its tone.