Command Line Interface¶

Sentimatrix includes a powerful CLI for quick analysis without writing code.

Installation¶

The CLI is installed automatically with Sentimatrix:

pip install sentimatrix

Verify installation:

sentimatrix --version

Basic Commands¶

Analyze Text¶

Analyze sentiment of text directly:

# Single text
sentimatrix analyze "This product is amazing!"

# Output
# Sentiment: positive
# Confidence: 96.7%
# Scores: positive=0.967, negative=0.021, neutral=0.012

Analyze from File¶

Analyze text from a file (one text per line):

sentimatrix analyze --file reviews.txt

With JSON output:

sentimatrix analyze --file reviews.txt --format json > results.json

Analyze with Emotions¶

Include emotion detection:

sentimatrix analyze "I'm so excited about this!" --emotions

# Output
# Sentiment: positive (97.2%)
# Emotions:
#   Primary: joy (89.3%)
#   Secondary: surprise (45.2%), anticipation (32.1%)

Scraping Commands¶

Scrape Reviews¶

Scrape reviews from supported platforms:

# Steam reviews
sentimatrix scrape https://store.steampowered.com/app/1245620 \
  --platform steam \
  --max-reviews 50

# Amazon reviews (requires browser)
sentimatrix scrape https://www.amazon.com/dp/B0BSHF7WHW \
  --platform amazon \
  --max-reviews 30 \
  --browser

Scrape and Analyze¶

Combine scraping with analysis:

sentimatrix scrape-analyze https://store.steampowered.com/app/1245620 \
  --platform steam \
  --max-reviews 100 \
  --output analysis.json

List Supported Platforms¶

sentimatrix platforms

# Output
# Platform Scrapers:
#   amazon     - Amazon product reviews
#   steam      - Steam game reviews
#   youtube    - YouTube comments
#   reddit     - Reddit posts and comments
#   imdb       - IMDB movie reviews
#   yelp       - Yelp business reviews
#   trustpilot - Trustpilot reviews
#   google     - Google Reviews

Configuration¶

Using Environment Variables¶

export GROQ_API_KEY="gsk_..."
export SENTIMATRIX_LLM_PROVIDER="groq"
export SENTIMATRIX_LLM_MODEL="llama-3.3-70b-versatile"

sentimatrix analyze "Test text" --llm

Using Config File¶

Create sentimatrix.yaml:

llm:
  provider: groq
  model: llama-3.3-70b-versatile

scraper:
  rate_limit:
    requests_per_second: 2

Use it:

sentimatrix analyze "Test" --config sentimatrix.yaml

Specify Config Path¶

sentimatrix --config /path/to/config.yaml analyze "Test text"

Output Formats¶

Text (Default)¶

sentimatrix analyze "Great product!"

# Sentiment: positive
# Confidence: 96.7%

JSON¶

sentimatrix analyze "Great product!" --format json

# {"sentiment": "positive", "confidence": 0.967, "scores": {...}}

CSV¶

sentimatrix analyze --file reviews.txt --format csv > results.csv

Table¶

sentimatrix analyze --file reviews.txt --format table

# ┌────────────────────────────────┬───────────┬────────────┐
# │ Text                           │ Sentiment │ Confidence │
# ├────────────────────────────────┼───────────┼────────────┤
# │ Great product!                 │ positive  │ 96.7%      │
# │ Terrible experience            │ negative  │ 94.2%      │
# │ It's okay                      │ neutral   │ 78.3%      │
# └────────────────────────────────┴───────────┴────────────┘

LLM Commands¶

Summarize Reviews¶

sentimatrix summarize --file reviews.txt --llm groq

# Summary:
# The product receives generally positive feedback with customers
# praising its build quality and value. Common complaints include
# shipping delays and limited color options.

Generate Insights¶

sentimatrix insights --file reviews.txt --llm groq

# Pros:
#   + Excellent build quality
#   + Great value for money
#   + Fast shipping (usually)
#
# Cons:
#   - Limited color options
#   - Occasional quality control issues
#
# Recommendation: Buy

Batch Processing¶

Process Directory¶

sentimatrix batch /path/to/reviews/ \
  --pattern "*.txt" \
  --output results/ \
  --format json

Parallel Processing¶

sentimatrix batch /path/to/reviews/ \
  --workers 4 \
  --output results/

Utility Commands¶

Check Configuration¶

sentimatrix config check

# Configuration:
#   LLM Provider: groq
#   LLM Model: llama-3.3-70b-versatile
#   Rate Limit: 2 req/s
#   Cache: memory
# Status: Valid

List Providers¶

sentimatrix providers

# Cloud Providers:
#   openai     - OpenAI (GPT-4o, GPT-4o-mini)
#   anthropic  - Anthropic (Claude 3.5 Sonnet)
#   google     - Google (Gemini 2.0, 1.5)
#   groq       - Groq (LLaMA 3.3, Mixtral)
#   mistral    - Mistral (7B, 8x7B, Large)
#   cohere     - Cohere (Command R+)
#
# Inference Providers:
#   together   - Together AI (200+ models)
#   fireworks  - Fireworks AI
#   openrouter - OpenRouter
#   cerebras   - Cerebras
#   deepseek   - DeepSeek
#
# Local Providers:
#   ollama     - Ollama (localhost:11434)
#   lmstudio   - LM Studio (localhost:1234)
#   vllm       - vLLM
#   llamacpp   - llama.cpp

Test Provider¶

sentimatrix test-provider groq

# Testing Groq...
# Model: llama-3.3-70b-versatile
# Status: Connected
# Latency: 142ms
# Test: PASSED

Command Reference¶

Command	Description
`analyze`	Analyze sentiment of text
`scrape`	Scrape reviews from a platform
`scrape-analyze`	Scrape and analyze in one step
`summarize`	Generate LLM summary
`insights`	Generate pros/cons insights
`batch`	Process multiple files
`platforms`	List supported platforms
`providers`	List LLM providers
`config`	Configuration management
`test-provider`	Test LLM provider connection

Global Options¶

Option	Description
`--config FILE`	Path to configuration file
`--format FORMAT`	Output format (text, json, csv, table)
`--verbose`	Enable verbose output
`--quiet`	Suppress all output except results
`--version`	Show version and exit
`--help`	Show help message

Examples¶

E-commerce Analysis¶

# Scrape and analyze Amazon product
sentimatrix scrape-analyze "https://www.amazon.com/dp/B0BSHF7WHW" \
  --platform amazon \
  --max-reviews 100 \
  --emotions \
  --llm groq \
  --output product_analysis.json

Gaming Review Dashboard¶

# Analyze multiple Steam games
for game_id in 1245620 1174180 292030; do
  sentimatrix scrape-analyze "https://store.steampowered.com/app/$game_id" \
    --platform steam \
    --max-reviews 50 \
    --output "game_${game_id}.json"
done

# Analyze Reddit discussion
sentimatrix scrape "https://reddit.com/r/technology/comments/abc123" \
  --platform reddit \
  --max-comments 200 \
  | sentimatrix analyze --stdin --emotions

Next Steps¶

Configuration - Configure the CLI
Examples - More CLI examples
API Reference - Python API documentation