Features Sentimatrix provides a comprehensive suite of text analysis features.
Feature Categories :material-text-box-search: Sentiment Analysis Multiple analysis modes from quick classification to aspect-based analysis.
Learn more →
:material-emoticon: Emotion Detection Detect emotions using multiple taxonomies with intensity analysis.
Learn more →
:material-image-multiple: Multi-Modal Analyze text, audio, images, and video for comprehensive understanding.
Learn more →
:material-brain: LLM Features Summarization, insight generation, and reasoning with 19 LLM providers.
Learn more →
Feature Overview Sentiment Analysis Feature Description Use Case Quick Sentiment Fast positive/negative/neutral High-volume processing Structured Analysis Detailed scores and confidence Analytics dashboards Fine-Grained 5-class classification Nuanced understanding Aspect-Based Per-aspect sentiment Product feedback Comparative Compare across items Competitor analysis Temporal Track changes over time Trend monitoring Domain-Specific Specialized for industry Finance, healthcare
Emotion Detection Feature Description Use Case Ekman Basic 6 universal emotions General use GoEmotions 28 fine-grained emotions Detailed analysis Plutchik Wheel 8 primary + combinations Psychological research Intensity Low/medium/high levels Customer satisfaction Timeline Emotion changes over text Narrative analysis
Multi-Modal Analysis Modality Capabilities Formats Text Full sentiment/emotion Any text Audio Speech-to-text + analysis WAV, MP3, FLAC Image Visual sentiment, OCR PNG, JPEG, WEBP Video Frame extraction + audio MP4, AVI, MOV
LLM -Powered Features Feature Description Requires LLM Summarization Generate review summaries Yes Insights Extract pros/cons/themes Yes Reasoning Chain-of-thought analysis Yes Comparison Compare products/services Yes
Quick Examples Sentiment Analysis async with Sentimatrix () as sm :
# Quick sentiment
result = await sm . analyze ( "Great product!" )
print ( f " { result . sentiment } : { result . confidence : .0% } " )
# Aspect-based
aspects = await sm . analyze_aspects (
"Great camera but battery life is poor" ,
aspects = [ "camera" , "battery" , "design" ]
)
# {'camera': 'positive', 'battery': 'negative', 'design': 'neutral'}
Emotion Detection async with Sentimatrix () as sm :
emotions = await sm . detect_emotions ( "I'm thrilled about this!" )
print ( f "Primary: { emotions . primary } " ) # joy
print ( f "Top 3: { emotions . top_k ( 3 ) } " ) # [('joy', 0.89), ('surprise', 0.12), ...]
LLM Featuresasync with Sentimatrix ( config ) as sm :
# Summarize reviews
summary = await sm . summarize_reviews ( reviews )
# Generate insights
insights = await sm . generate_insights ( reviews )
print ( f "Pros: { insights . pros } " )
print ( f "Cons: { insights . cons } " )
Sentiment Analysis Model Accuracy Latency (CPU ) Latency (GPU ) RoBERTa-base 94.8% 50ms 9ms DistilBERT 92.1% 25ms 5ms BERT-base 93.5% 60ms 12ms
Emotion Detection Model Accuracy Classes Latency GoEmotions 65.2% 28 45ms Ekman 78.3% 6 35ms
Throughput Mode Batch Size Throughput Single 1 20/sec Batch 32 400/sec Batch (GPU ) 32 2000/sec
Feature Documentation Sentiment Analysis Emotion Detection Multi-Modal LLM Features