GitHub - fshasan/feedback-pulse: An AI-powered product feedback aggregation and analysis tool built on Cloudflare Workers. Automatically analyzes customer feedback from multiple sources using Llama AI models to extract sentiment, themes, urgency, and actionable insights.

🎯 Overview

Feedback Pulse helps product teams aggregate, analyze, and prioritize feedback from various channels (Support, Discord, GitHub, Twitter, Email, Forums) using advanced AI analysis. The tool provides contextual understanding of feedback beyond simple keyword matching, enabling data-driven product decisions.

Feedback Input Approach: Instead of integrating with third-party APIs, Feedback Pulse uses a manual input system where users can enter feedback through a text field and select the source via a dropdown menu. This approach provides flexibility and control, allowing users to analyze feedback from any source without requiring API keys, webhooks, or complex integrations. Simply paste or type the feedback content, select the appropriate source channel, and click "Analyze" to get instant AI-powered insights.

Assignment Compliance: This prototype meets all requirements for the Cloudflare Product Manager Intern Assignment:

✅ Deployed on Cloudflare Workers
✅ Uses 3 Cloudflare Developer Platform products (Workers, D1, Workers AI)
✅ Uses mock data (no third-party API integrations required)
✅ Includes architecture overview explaining product choices and rationale
✅ Extracts themes, urgency, value, and sentiment from feedback
✅ Provides interactive dashboard for aggregated feedback analysis

✨ Features

🤖 AI-Powered Analysis

Contextual Sentiment Analysis: Uses Llama 3.1 70B to understand sentiment beyond simple keywords, analyzing tone and context
Smart Theme Extraction: Automatically identifies topics like Performance, API, Security, Documentation, Bugs, Reliability, Features, Pricing
Value & Urgency Scoring: AI-driven prioritization (0-10 scale) based on business impact, source credibility, and criticality
Spam & Off-topic Detection: AI-powered detection of spam, promotional content, and off-topic messages. Spam is stored in database for moderation but analysis is skipped
Meaningless Content Detection: Filters out gibberish and non-substantive input using AI validation
Detailed Explanations: AI-generated reasoning for each analysis decision with contextual insights

📊 Interactive Dashboard

Manual Feedback Input: Simple text field and source dropdown for entering feedback - no third-party API integrations required. Paste feedback from any source and analyze instantly.
Real-time Feedback Submission: Add and analyze new feedback instantly with immediate AI processing
Advanced Filtering: Filter by source, sentiment, or search by keywords with real-time results
Sortable Tables: Sort by date, sentiment, value score, or urgency with visual indicators
Pagination: Navigate through large datasets efficiently (5, 10, 25, or 50 items per page)
Click-to-View Analysis: Select any feedback item to see detailed analysis with color-coded metrics
Responsive Design: Works seamlessly on desktop, tablet, and mobile devices

💾 Persistent Storage

Cloudflare D1 Database: All feedback is stored persistently in SQLite database
Survives Server Restarts: Data remains intact across deployments and restarts
Efficient Queries: Indexed columns for fast filtering, sorting, and searching
Fallback Support: Gracefully falls back to in-memory storage if database is unavailable

🎯 Supported Sources

Customer Support - Support tickets and inquiries
Discord - Community messages and discussions
GitHub - Issues, feature requests, and bug reports
Twitter/X - Social media posts and mentions
Email - Direct email feedback
Community Forum - Forum posts and discussions

🏗️ Architecture

Cloudflare Products Used

This prototype uses 3 Cloudflare Developer Platform products:

Cloudflare Workers (Runtime)
- Why: Serverless edge computing platform that provides global distribution, zero cold starts, and automatic scaling. Perfect for building APIs and serving static assets without managing infrastructure. Workers handle all backend logic, API endpoints, and serve the frontend dashboard.
Cloudflare D1 (Database)
- Why: Serverless SQLite database that provides persistent storage without database management overhead. D1 integrates seamlessly with Workers, offers automatic backups, and scales automatically. Essential for storing feedback data persistently across deployments and server restarts.
Cloudflare Workers AI (AI Inference)
- Why: Edge AI platform that runs Llama models directly on Cloudflare's network, eliminating the need for external AI APIs or API keys. Provides low-latency AI inference, no rate limits, and integrates natively with Workers. Enables contextual analysis using Meta's Llama 3.1 models without additional infrastructure.

Technology Stack

Component	Technology	Purpose
Runtime	Cloudflare Workers	Serverless edge computing platform
Database	Cloudflare D1 (SQLite)	Persistent storage for feedback data
AI Models	Cloudflare Workers AI	Llama 3.1 models for analysis
Frontend	Vanilla HTML/CSS/JavaScript	Lightweight, fast-loading dashboard
API	RESTful endpoints	JSON-based API for data operations
Deployment	Wrangler CLI	Cloudflare deployment tool

AI Models Used

Model	Purpose	Fallback
`@cf/meta/llama-3.1-70b-instruct`	Primary analysis & explanations	8B model
`@cf/meta/llama-3.1-8b-instruct`	Validation & fallback	Rule-based

System Flow

User Input → AI Validation → Spam Check → AI Analysis → Database Storage → Dashboard Display
     ↓              ↓              ↓            ↓              ↓                ↓
  Feedback    Meaningless?    Spam?      Sentiment      D1 SQLite      Real-time UI
   Form       Detection       Skip       Themes         Storage         Updates
                            Analysis    Urgency/Value   (All records)

Spam Handling: If spam/off-topic content is detected, analysis is skipped but the record is still stored in the database with a spam flag for moderation purposes.

📋 Prerequisites

Node.js (v18 or higher)
npm (comes with Node.js)
Cloudflare Account (for deployment)
Wrangler CLI (installed via npm)

🚀 Getting Started

Installation

Clone the repository

git clone https://github.com/fshasan/feedback-pulse.git
cd feedback-pulse

Install dependencies

npm install

Local Development

Set up local D1 database

npx wrangler d1 migrations apply feedback-db --local

Note: If you need to reset the database (clear all data and reapply migrations):

# Drop existing tables
npx wrangler d1 execute feedback-db --local --command "DROP TABLE IF EXISTS feedback"
npx wrangler d1 execute feedback-db --local --command "DROP TABLE IF EXISTS d1_migrations"

# Reapply migrations
npx wrangler d1 migrations apply feedback-db --local

Or use the reset script:

./scripts/reset-db.sh

Start the development server

npm run dev

Open your browser

http://localhost:8787

The development server will automatically reload when you make changes to the code.

Tip: Check out EXAMPLES.md for sample feedback messages you can use to test the system, including valid feedback, spam examples, and edge cases.

Production Deployment

Create D1 database (first time only)

npx wrangler d1 create feedback-db

Update wrangler.jsonc with your database ID
- Copy the database_id from the command output
- Update the database_id field in wrangler.jsonc
Apply migrations to remote database

npx wrangler d1 migrations apply feedback-db --remote

Note: If you need to reset the remote database:

# Drop existing tables
npx wrangler d1 execute feedback-db --remote --command "DROP TABLE IF EXISTS feedback"
npx wrangler d1 execute feedback-db --remote --command "DROP TABLE IF EXISTS d1_migrations"

# Reapply migrations
npx wrangler d1 migrations apply feedback-db --remote

Deploy to Cloudflare Workers

npx wrangler deploy

Your application will be available at: https://feedback-pulse.<your-subdomain>.workers.dev

📡 API Documentation

Base URL

Production: https://feedback-pulse.<your-subdomain>.workers.dev
Local: http://localhost:8787

Endpoints

GET `/api/feedback`

Retrieve paginated feedback with optional filters and sorting.

Query Parameters:

Parameter	Type	Description	Default	Example
`page`	number	Page number (1-indexed)	`1`	`?page=2`
`limit`	number	Items per page	`10`	`?limit=25`
`source`	string	Filter by source	`all`	`?source=discord`
`sentiment`	string	Filter by sentiment	`all`	`?sentiment=negative`
`search`	string	Search in content/title	-	`?search=bug`
`sortBy`	string	Sort column	`timestamp`	`?sortBy=urgencyScore`
`sortOrder`	string	Sort direction (`asc`/`desc`)	`desc`	`?sortOrder=asc`

Response:

{
  "data": [
    {
      "id": "CS-001",
      "source": "support",
      "title": "API rate limiting too restrictive",
      "content": "We keep hitting rate limits...",
      "author": "sarah@company.com",
      "timestamp": "2025-01-20T10:00:00.000Z",
      "metadata": { "priority": "high" },
      "analysis": {
        "sentiment": "neutral",
        "sentimentScore": 0,
        "themes": ["API", "Features"],
        "valueScore": 6.0,
        "urgencyScore": 8.0,
        "isMeaningless": false,
        "reasoning": "AI-generated explanation..."
      }
    }
  ],
  "pagination": {
    "page": 1,
    "limit": 10,
    "totalCount": 25,
    "totalPages": 3,
    "hasNext": true,
    "hasPrev": false
  }
}

POST `/api/feedback/save`

Save new feedback to the database.

Request Body:

{
  "id": "NEW-1234567890",
  "source": "discord",
  "title": "Feature request",
  "content": "Would love to see dark mode support",
  "author": "user#1234",
  "timestamp": "2025-01-20T12:00:00.000Z",
  "metadata": {},
  "analysis": {
    "sentiment": "positive",
    "sentimentScore": 2,
    "themes": ["Features"],
    "valueScore": 5.5,
    "urgencyScore": 2,
    "isMeaningless": false,
    "reasoning": "..."
  }
}

Response:

{
  "success": true,
  "message": "Feedback saved to database",
  "id": "NEW-1234567890"
}

POST `/api/ai/validate`

Validate if feedback content is meaningful, spam, or offensive.

Request Body:

{
  "content": "This is meaningful feedback about the product"
}

Response:

{
  "isMeaningless": false,
  "isSpam": false,
  "reason": "Content appears to be meaningful feedback"
}

Note: If isSpam is true, the content will be stored in the database but no analysis will be performed.

POST `/api/ai/analyze`

Get comprehensive AI analysis for feedback.

Request Body:

{
  "content": "The API is too slow and causing timeouts",
  "source": "github",
  "metadata": { "priority": "high" }
}

Response:

{
  "sentiment": "negative",
  "sentimentScore": -2,
  "themes": ["Performance", "API", "Reliability"],
  "valueScore": 7.5,
  "urgencyScore": 8,
  "isUrgent": true,
  "isMeaningless": false,
  "reasoning": "The feedback expresses frustration with API performance..."
}

POST `/api/ai/explain`

Get detailed AI explanations for analysis metrics.

Request Body:

{
  "feedback": { /* feedback object */ },
  "analysis": { /* analysis object */ }
}

Response:

{
  "sentiment": "Explanation of sentiment analysis...",
  "valueScore": "Explanation of value score...",
  "urgencyScore": "Explanation of urgency score...",
  "themes": "Explanation of theme extraction...",
  "summary": "Overall assessment..."
}

🗄️ Database Schema

`feedback` Table

Column	Type	Description
`id`	TEXT (PK)	Unique identifier
`source`	TEXT	Feedback source (support, discord, github, etc.)
`title`	TEXT	Feedback title/summary
`content`	TEXT	Full feedback content
`author`	TEXT	Author identifier
`timestamp`	TEXT	ISO timestamp
`metadata`	TEXT (JSON)	Additional metadata
`sentiment`	TEXT	Sentiment classification
`sentiment_score`	REAL	Sentiment score (-3 to +3)
`themes`	TEXT (JSON)	Array of identified themes
`value_score`	REAL	Value score (0-10)
`urgency_score`	REAL	Urgency score (0-10)
`is_meaningless`	INTEGER	Boolean flag (0/1)
`is_spam`	INTEGER	Boolean flag (0/1) - spam/off-topic content
`reasoning`	TEXT	AI-generated reasoning
`created_at`	TEXT	Creation timestamp

Indexes:

idx_feedback_source on source
idx_feedback_timestamp on timestamp
idx_feedback_sentiment on sentiment
idx_feedback_is_spam on is_spam

📁 Project Structure

feedback-pulse/
├── src/
│   └── index.js                    # Cloudflare Worker backend
│                                    # - API endpoints
│                                    # - AI integration
│                                    # - Database operations
│                                    # - Static file serving
├── public/
│   └── index.html                  # Frontend dashboard
│                                    # - Interactive UI
│                                    # - API client
│                                    # - Real-time updates
├── migrations/
│   ├── 0001_create_feedback_table.sql  # Database schema (includes is_spam)
│   └── 0002_add_is_spam_column.sql     # Migration for existing databases
├── scripts/
│   ├── setup-db.sh                 # Database setup helper
│   └── reset-db.sh                # Database reset script
├── test/
│   └── index.spec.js                # Unit tests
├── wrangler.jsonc                  # Cloudflare configuration
├── package.json                     # Dependencies & scripts
├── vitest.config.js                # Test configuration
├── README.md                        # This file
├── EXAMPLES.md                      # Feedback examples for testing
└── FRICTION_LOG.md                  # Development friction log

🔧 Configuration

`wrangler.jsonc`

Key configuration options:

{
  "name": "feedback-pulse",              // Worker name
  "main": "src/index.js",                // Entry point
  "compatibility_date": "2025-09-27",   // Compatibility date
  "compatibility_flags": [
    "nodejs_compat",                     // Node.js compatibility
    "global_fetch_strictly_public"       // Global fetch
  ],
  "assets": {
    "directory": "./public"              // Static assets directory
  },
  "ai": {
    "binding": "AI"                      // AI binding name
  },
  "d1_databases": [
    {
      "binding": "DB",                   // Database binding name
      "database_name": "feedback-db",    // Database name
      "database_id": "..."               // Database UUID
    }
  ]
}

🧪 Testing

Run the test suite:

npm test

Testing Feedback Examples

See EXAMPLES.md for a comprehensive collection of feedback examples including:

Valid Feedback:

Positive feedback (appreciation, success stories)
Negative feedback (bugs, issues, complaints)
Neutral feedback (feature requests, questions)
High urgency feedback (critical issues, outages)
Performance, security, and documentation feedback

Spam Examples:

Promotional/advertising spam
Off-topic/irrelevant content
Cryptocurrency/investment scams
Offensive/inappropriate content
Phishing attempts

Meaningless Content:

Repetitive patterns
Random character sequences
Gibberish text

Use Cases:

Test spam detection accuracy
Validate analysis quality across different feedback types
Test edge cases and boundary conditions
Demonstrate system capabilities
Training and documentation purposes

Simply copy and paste examples from EXAMPLES.md into the feedback form to test different scenarios.

🐛 Troubleshooting

Database Issues

Problem: binding DB of type d1 must have a valid 'id' specified

Solution:

Create database: npx wrangler d1 create feedback-db
Copy the database_id from output
Update wrangler.jsonc with the correct database_id
Restart dev server

Problem: Data not persisting

Solution:

Ensure migrations are applied: npx wrangler d1 migrations apply feedback-db --remote
Check database binding in wrangler.jsonc
Verify database ID is correct

Problem: table feedback has no column named is_spam

Solution:

Apply the latest migration: npx wrangler d1 migrations apply feedback-db --local (or --remote)
Or reset database and reapply all migrations using ./scripts/reset-db.sh

AI Analysis Issues

Problem: AI analysis not working

Solution:

Verify AI binding is configured in wrangler.jsonc
Check Cloudflare Workers AI is enabled in your account
The system will automatically fallback to rule-based analysis if AI is unavailable

Deployment Issues

Problem: Changes not reflecting after deployment

Solution:

Clear browser cache
Hard refresh (Ctrl+Shift+R / Cmd+Shift+R)
Verify deployment succeeded: npx wrangler deployments list

🔐 Security Considerations

CORS: Currently allows all origins (*) - restrict in production
Input Validation: All user input is validated before processing
SQL Injection: Uses parameterized queries via D1 prepared statements
AI Rate Limiting: Consider implementing rate limits for AI endpoints
Authentication: Add authentication for production deployments

🚧 Future Enhancements

📝 License

MIT License - feel free to use this project for your own purposes.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Fork the repository
Create your feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add some amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

📧 Support

For issues, questions, or contributions, please open an issue on GitHub.

Built with ❤️ using Cloudflare Workers, Workers AI, and D1

Name		Name	Last commit message	Last commit date
Latest commit History 12 Commits
.vscode		.vscode
migrations		migrations
public		public
scripts		scripts
src		src
test		test
.editorconfig		.editorconfig
.gitignore		.gitignore
.prettierrc		.prettierrc
EXAMPLES.md		EXAMPLES.md
FRICTION_LOG.md		FRICTION_LOG.md
README.md		README.md
package-lock.json		package-lock.json
package.json		package.json
vitest.config.js		vitest.config.js
wrangler.jsonc		wrangler.jsonc

Folders and files

Latest commit

History

Repository files navigation

🎯 Overview

✨ Features

🤖 AI-Powered Analysis

📊 Interactive Dashboard

💾 Persistent Storage

🎯 Supported Sources

🏗️ Architecture

Cloudflare Products Used

Technology Stack

AI Models Used

System Flow

📋 Prerequisites

🚀 Getting Started

Installation

Local Development

Production Deployment

📡 API Documentation

Base URL

Endpoints

GET /api/feedback

POST /api/feedback/save

POST /api/ai/validate

POST /api/ai/analyze

POST /api/ai/explain

🗄️ Database Schema

feedback Table

📁 Project Structure

🔧 Configuration

wrangler.jsonc

🧪 Testing

Testing Feedback Examples

🐛 Troubleshooting

Database Issues

AI Analysis Issues

Deployment Issues

🔐 Security Considerations

🚧 Future Enhancements

📝 License

🤝 Contributing

📧 Support

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

GET `/api/feedback`

POST `/api/feedback/save`

POST `/api/ai/validate`

POST `/api/ai/analyze`

POST `/api/ai/explain`

`feedback` Table

`wrangler.jsonc`

Packages