Product Research

from pyagi import AGIClient

client = AGIClient(api_key="your_api_key")

with client.session("agi-0") as session:
    result = session.run_task(f"""
        Complete research report for: {product_name}

    Section 1: Product Overview
    - Full product name and model number
    - Brand and manufacturer
    - Category and subcategory
    - Release date / launch year
    - Current availability status

    Section 2: Specifications
    - All technical specifications
    - Dimensions and weight
    - Color/size options available
    - Materials and build quality
    - Warranty information

    Section 3: Pricing Analysis
    - Current price range across retailers
    - MSRP / list price
    - Historical price trends (if available)
    - Current promotions or discounts
    - Best price currently available (with retailer)

    Section 4: Reviews & Ratings
    - Overall rating average (across platforms)
    - Total number of reviews
    - Rating distribution (5 star: X%, 4 star: Y%, etc.)
    - Top 5 positive themes from reviews
    - Top 5 negative themes from reviews
    - Expert review scores (if available)

    Section 5: Comparison & Alternatives
    - Top 3 direct competitors
    - Key differentiating features
    - Price comparison with competitors
    - Why users choose this vs alternatives

    Section 6: Buying Recommendations
    - Best retailer to buy from (price + shipping + reliability)
    - Best time to buy (based on pricing trends)
    - Who this product is best for
    - Who should consider alternatives

    Return as structured JSON with all sections and sources cited.
    """)

export AGI_API_KEY="your_api_key"

SESSION=$(http POST https://api.agi.tech/v1/sessions \
  Authorization:"Bearer $AGI_API_KEY" \
  agent_name=agi-0 | jq -r '.session_id')

http POST https://api.agi.tech/v1/sessions/$SESSION/message \
  Authorization:"Bearer $AGI_API_KEY" \
  message="Complete research report for: $product_name. Include: Product Overview, Specifications, Pricing Analysis, Reviews & Ratings, Comparison & Alternatives, Buying Recommendations. Return as structured JSON with all sections and sources cited."

# Poll for results
while true; do
  STATUS=$(http GET https://api.agi.tech/v1/sessions/$SESSION/status \
    Authorization:"Bearer $AGI_API_KEY" | jq -r '.status')
  if [ "$STATUS" = "finished" ]; then
    http GET https://api.agi.tech/v1/sessions/$SESSION/messages \
      Authorization:"Bearer $AGI_API_KEY" | jq '.messages[] | select(.type=="DONE")'
    break
  fi
  sleep 2
done

# Cleanup
http DELETE https://api.agi.tech/v1/sessions/$SESSION \
  Authorization:"Bearer $AGI_API_KEY"

result = session.run_task(f"""
    Analyze customer reviews for product at: {product_url}

    Collect at least {min_reviews} reviews if available.

    Analysis Required:
    1. Quantitative Metrics: Overall rating, total review count, rating distribution, verified purchase percentage, review recency
    2. Sentiment Analysis: Overall sentiment, sentiment by rating level, sentiment trends over time
    3. Theme Extraction: Top themes in positive/negative reviews
    4. Feature Mentions: Build quality, performance, value, customer service, ease of use, durability, design
    5. Use Cases: How customers use the product, unexpected use cases, demographics
    6. Comparison Mentions: Products compared to, upgrade/downgrade patterns, cross-shopping behavior

    Return as JSON with summary, positive_themes, negative_themes, feature_sentiment, and recommendations.
""")

http POST https://api.agi.tech/v1/sessions/$SESSION/message \
  Authorization:"Bearer $AGI_API_KEY" \
  message="Analyze customer reviews for product at: $product_url. Collect at least $min_reviews reviews. Return JSON with quantitative metrics, sentiment analysis, theme extraction, feature mentions, use cases, and comparison mentions."

products = ['Product A', 'Product B', 'Product C']
result = session.run_task(f"""
    Compare specifications for: {', '.join(products)}

    For each product, extract: All technical specifications, dimensions, weight, materials, features, compatibility, warranty, price.

    Create side-by-side comparison table with: Feature comparison, specification differences, unique features per product, price comparison, value assessment.

    Analysis: Best overall, best for specific use cases, best value, trade-offs.

    Return as JSON with comparison matrix and recommendations.
""")

http POST https://api.agi.tech/v1/sessions/$SESSION/message \
  Authorization:"Bearer $AGI_API_KEY" \
  message="Compare specifications for: Product A, Product B, Product C. Extract all technical specs, dimensions, features, price for each. Create side-by-side comparison with feature differences, unique features, value assessment. Return JSON with comparison matrix and recommendations."

result = session.run_task(f"""
    Find and summarize expert reviews for: {product_name}

    Sources: Major tech review sites (CNET, TechCrunch, The Verge), specialized reviewers, YouTube video reviews, Consumer Reports.

    For each review, extract: Publication/reviewer name, review date, overall score (0-10), pros, cons, key quotes, verdict, review URL.

    Analysis: Average expert rating, consensus pros/cons, unique insights, awards.

    Synthesis: What experts agree/disagree on, overall consensus, concerns.

    Return as JSON with individual reviews and synthesis.
""")

http POST https://api.agi.tech/v1/sessions/$SESSION/message \
  Authorization:"Bearer $AGI_API_KEY" \
  message="Find and summarize expert reviews for: $product_name. Check major tech review sites, YouTube reviews, Consumer Reports. Extract publication, date, score, pros, cons, quotes, verdict, URL. Return JSON with individual reviews and synthesis."

retailers = ['Amazon', 'Best Buy', 'Target']
result = session.run_task(f"""
    Analyze price history for: {product_name}

    Retailers: {', '.join(retailers)}

    Find: Historical Pricing (current, lowest/highest in 6 months, average, launch price), Price Patterns (seasonal trends, sale frequency, discount %, day/week patterns), Prediction (current price assessment, expected direction, best time to buy, next sale date), Retailer Analysis (best price retailer, sale frequency, price matching, shipping).

    Return as JSON with price history data, visual data for charting, buying recommendation, and price alert suggestion.
""")

http POST https://api.agi.tech/v1/sessions/$SESSION/message \
  Authorization:"Bearer $AGI_API_KEY" \
  message="Analyze price history for: $product_name. Retailers: Amazon, Best Buy, Target. Find historical pricing, price patterns, predictions, and retailer analysis. Return JSON with price history data, visual data, buying recommendation, and price alert suggestion."

location_str = f" near {location}" if location else ""
result = session.run_task(f"""
    Check availability for: {product_name}{location_str}

    Retailers: Amazon, Best Buy, Target, Walmart, Apple Store, Manufacturer website.

    For each retailer: Availability Status (in stock online/in-store, delivery date, pickup), Variants (available/out of stock colors/sizes/models, restrictions), Pricing (current price, extra costs, total cost), Alerts (restock alerts, notifications, back-order).

    Summary: Where to buy immediately, fastest delivery, cheapest option, best overall.

    Return as JSON with availability matrix.
""")

http POST https://api.agi.tech/v1/sessions/$SESSION/message \
  Authorization:"Bearer $AGI_API_KEY" \
  message="Check availability for: $product_name. Retailers: Amazon, Best Buy, Target, Walmart, Apple Store, Manufacturer website. For each retailer provide availability status, variants, pricing, and alerts. Return JSON with availability matrix."

result = session.run_task(f"""
    Find and analyze unboxing content for: {product_name}

    Sources: YouTube unboxing videos (top 10-20), blog posts, social media first impressions, forum posts.

    Extract: Packaging Quality (design feedback, experience rating, accessories, setup difficulty), First Impressions (initial reactions, build quality, size/weight, vs expectations), Common Themes (surprises, concerns, excitement, disappointments), Setup Experience (time, difficulty, common issues, instructions quality).

    Summarize: Overall sentiment, key takeaways, red flags, standout positives.

    Return as JSON with themes and examples.
""")

http POST https://api.agi.tech/v1/sessions/$SESSION/message \
  Authorization:"Bearer $AGI_API_KEY" \
  message="Find and analyze unboxing content for: $product_name. Sources: YouTube, blogs, social media, forums. Extract packaging quality, first impressions, common themes, setup experience. Return JSON with themes and examples."