Test any text against all sentiment models in real-time
0 / 5000
Word-Level Analysis
Statement Analyzer
Score and archive statements for sentiment tracking over time
0 / 5000
Loading statement history...
Lexicon Browser
Search and explore all sentiment model vocabularies
Detection Capabilities
Current and planned analysis engines
Aa
Unigram Lexicon Scoring
Single-word sentiment detection across 4 specialized models. 800+ curated words with data-validated weights from Twitter and Reddit corpora.
ACTIVE
Ab
Bigram & Phrase Detection
Two-word combinations like "peace deal", "civil war", "arms control" that carry stronger signal than individual words. 100+ phrases across models.
ACTIVE
ML
AWS Comprehend (NLP)
Machine learning sentiment analysis via Amazon Comprehend. Contextual understanding beyond keyword matching. Enriches top 50 trending topics.
ACTIVE
Tr
Auto-Translation (60+ Languages)
Automatic detection and translation of non-English content via AWS Translate. SigV4 signed requests with language detection and smart filtering.
ACTIVE
Img
Image Sentiment Analysis
Detect sentiment from images using visual cues — facial expressions, color palettes, scene composition, text-in-image extraction (OCR). Planned integration with AWS Rekognition.
PLANNED
Em
Emoji & Symbol Detection
Map emojis, emoticons, and Unicode symbols to sentiment values. Social media posts use emojis as primary sentiment indicators that pure text models miss.
PLANNED
Ctx
Contextual Negation
Handle negation patterns ("not good", "never happy", "far from ideal") that flip word-level sentiment. Uses n-gram context windows to detect polarity shifts.
PLANNED
Sar
Sarcasm Detection
Identify sarcastic and ironic statements where surface-level sentiment is inverted. Uses contrast patterns, hyperbole markers, and contextual cues.
RESEARCH
Training Data Sources
Public datasets used to train and validate sentiment models