ProduckAI MCP Server
Transform scattered voice of customer feedback into actionable insights using AI-powered analysis and seamless Claude Desktop integration.
🌟 What is ProduckAI?
ProduckAI MCP Server brings enterprise-grade product feedback analysis directly into your AI workflows. Seamlessly integrate with Claude Desktop to analyze customer feedback, generate insights, prioritize features, and create executive-ready PRDs—all using natural language.
Key Value Proposition
- 70% faster Scattered VOC -> AI-powered generation
- Multi-source ingestion: Slack, Google Drive, Zoom, JIRA, CSV
- Smart prioritization: 6-dimension VOC scoring
- Evidence-backed decisions: Every PRD linked to customer quotes
- 50 specialized tools: Complete workflow from collection to execution
Why Open Source?
Product management is evolving with AI, but most tools are closed source and expensive. This project exists to democratize AI-powered product management for teams of all sizes.
By open sourcing this MCP server, we're creating a platform where:
- 🤝 PMs share learnings - Your feedback analysis insights help others prioritize better
- 🔧 Engineers build together - Improve clustering algorithms, add integrations, enhance AI prompts
- 📚 Community grows knowledge - Document best practices, share PRD templates, teach new PMs
This isn't just a tool—it's a movement to make product management more data-driven, evidence-based, and accessible to everyone.
If you're a PM who's ever struggled with feedback overload, or an engineer building tools for PMs, this project is for you. Contribute,learn, and help shape the future of AI-assisted product management.
✨ Features
📥 Multi-Source Feedback Collection
- Slack - Auto-sync channels with AI-powered customer detection
- Google Drive - Process docs, PDFs with OCR
- Zoom - Auto-fetch recordings, AI transcript analysis
- JIRA - Bidirectional sync with issue tracking
- CSV/Manual - Bulk upload or quick capture
🧠 AI-Powered Analysis
- Semantic Clustering - Group similar feedback automatically
- Insight Generation - AI creates actionable themes
- Sentiment Detection - Identify urgent vs nice-to-have
- Customer Attribution - Auto-match feedback to customers
🎯 Smart Prioritization
- VOC Scoring - 6-dimension scoring (0-100):
- Customer Impact (30%) - Tier, revenue, strategic
- Frequency (20%) - How often mentioned
- Recency (15%) - How recent
- Sentiment (15%) - Urgency level
- Theme Alignment (10%) - Strategic fit
- Effort (10%) - Implementation complexity
📄 PRD Generation
- AI-Powered PRDs - Strategic documents from insights
- Evidence-Based - Includes direct customer quotes
- Segment-Aware - Tailored for Enterprise vs SMB
- Risk Assessment - Effort-based implementation risks
- Version Tracking - PRD history and updates
🔄 JIRA Integration
- Bidirectional Sync - Issues ↔ Feedback
- Auto-Priority - VOC score → JIRA priority
- Issue Creation - Generate epics from insights
- Linkage Tracking - Trace feedback to issues
🚀 Quick Start
Prerequisites
- Python 3.11+
- Anthropic API Key (get one here)
- Claude Desktop (download)
Installation
pip install produckai-mcp-server
Configuration
-
Create environment file:
cp .env.example .env # Edit .env and add your ANTHROPIC_API_KEY -
Configure Claude Desktop (
~/Library/Application Support/Claude/claude_desktop_config.json):{ "mcpServers": { "produckai": { "command": "produckai-mcp", "env": { "ANTHROPIC_API_KEY": "your-api-key-here" } } } } -
Restart Claude Desktop
First Use
Try these commands in Claude:
"Upload the demo feedback CSV at ./demo-data/feedback.csv"
"Run clustering and show me the top themes"
"Calculate VOC scores and show top 5 insights"
"Generate a PRD for the highest-priority insight"
🛠️ Available Tools (50 Total)
📥 Ingestion (21 tools)
- Slack: setup, sync channels, tag customers, bot filters
- Google Drive: setup, browse, sync folders, preview, processing config
- Zoom: setup, sync recordings, analyze meetings, insights, customer linking
- JIRA: setup, browse projects, bidirectional sync, mapping, reports
- Manual: CSV upload, Zoom transcript, raw capture, templates
⚙️ Processing (4 tools)
run_clustering- Generate themes and insightsgenerate_embeddings- Create vector embeddingsget_themes- List all themesget_theme_details- Deep-dive on theme
🔍 Query (4 tools)
search_insights- Natural language searchget_insight_details- Full insight datasearch_feedback- Search raw feedbackget_customer_feedback- Customer-specific view
🎯 VOC Scoring (4 tools)
calculate_voc_scores- Score feedback/themesget_top_feedback_by_voc- Priority-ranked listconfigure_voc_weights- Customize algorithmget_voc_trends- Track changes over time
📄 PRD Generation (6 tools)
generate_prd- Create PRD from insightlist_prds- Browse generated PRDsget_prd- View full PRDupdate_prd_status- Workflow trackingregenerate_prd- Update after changesexport_prd- Export to markdown
🏥 Management (11 tools)
- Status checks, sync monitoring, health checks, configuration
📊 Quick Reference
| Integration | Setup Time | Cost/Month | Key Features |
|---|---|---|---|
| Slack | 10 min | $1-2 | AI classification, delta sync, bot filtering |
| Google Drive | 15 min | $5-10 | Multi-format, comments, auto-detect |
| Zoom | 10 min | $3-4 | Auto-download, AI analysis, sentiment |
| JIRA | 5 min | Free | Bidirectional, VOC priority, evidence |
| CSV | 0 min | Free | Bulk upload, templates, quick capture |
| Feature | Time | Cost | Output |
|---|---|---|---|
| Clustering (100 items) | 1-2 min | $0.20 | Themes & insights |
| VOC Scoring (100 items) | 10 sec | $0.01 | Priority ranking (0-100) |
| PRD Generation | 10-15 sec | $0.05-0.10 | Strategic document |
📖 Complete Workflow Example
Weekly Feedback Triage (20 minutes)
# Monday: Collect feedback
"Sync Slack #customer-feedback channel for the last 7 days"
"Sync Zoom recordings from the past week"
"Upload the quarterly feedback CSV"
# Tuesday: Analyze
"Run clustering to identify themes"
"Show me the top 10 themes by feedback count"
# Wednesday: Prioritize
"Calculate VOC scores for all insights"
"Show me the top 5 highest-priority insights"
# Thursday: Document
"Generate a PRD for the top insight about API rate limiting"
"Export the PRD to ~/Documents/PRDs/"
# Friday: Execute
"Sync the top 3 insights to JIRA project PROD"
"Show JIRA sync status"
Result: 3 executive-ready PRDs, synced to JIRA, evidence-backed by customer feedback.
🏗️ Architecture
┌─────────────────┐
│ Data Sources │ Slack, Drive, Zoom, JIRA, CSV
└────────┬────────┘
│
▼
┌─────────────────┐
│ MCP Server │ 50 tools, state management, AI classification
│ (This Package) │
└────────┬────────┘
│
▼
┌─────────────────┐
│ ProduckAI API │ Clustering, insights, embeddings
│ (Optional) │
└────────┬────────┘
│
▼
┌─────────────────┐
│ Claude Desktop │ Natural language interface
└─────────────────┘
Deployment Model: Local single-user (each PM runs their own instance)
🧪 Demo Data
Try ProduckAI with sample data:
# Generate demo data (50 feedback items)
python scripts/generate_demo_data.py
# In Claude:
"Upload demo-data/feedback.csv"
"Run clustering"
"Generate PRD for top insight"
See demo-data/README.md for details.
🔐 Integration Setup
Slack
- Create Slack App: https://api.slack.com/apps
- Add scopes:
channels:history,channels:read,users:read - Install to workspace
- In Claude:
"Setup Slack integration"
Google Drive
- Create GCP project: https://console.cloud.google.com
- Enable Google Drive API
- Create OAuth credentials (Desktop app)
- In Claude:
"Setup Google Drive integration"
JIRA
- Generate API token: https://id.atlassian.com/manage/api-tokens
- In Claude:
"Setup JIRA integration with server URL, email, and token"
Zoom
- Create OAuth app: https://marketplace.zoom.us/develop/create
- Add scope:
recording:read:admin - In Claude:
"Setup Zoom integration"
See docs/ for detailed setup guides.
📚 Documentation
- Installation Guide
- Quick Start Guide
- Integration Setup
- End-to-End Workflow
- Open Source Roadmap
- Phase Implementation Docs
- PRD Generation Prompt
- Contributing Guide
🧑💻 Development
Setup
# Clone repository
git clone https://github.com/produckai/produckai-mcp-server.git
cd produckai-mcp-server
# Create virtual environment
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
# Install in development mode
pip install -e ".[dev]"
Testing
# Run tests
pytest
# Run with coverage
pytest --cov
# Run linting
ruff check .
# Format code
black .
# Type checking
mypy src/
Code Quality
We use:
- Black for code formatting
- Ruff for linting
- MyPy for type checking
- Pytest for testing
See CONTRIBUTING.md for guidelines.
🐛 Troubleshooting
MCP Server Not Appearing in Claude
- Check config:
cat ~/Library/Application\ Support/Claude/claude_desktop_config.json - Verify command:
which produckai-mcp - Check logs:
tail -f ~/.produckai/logs/mcp-server.log - Restart Claude Desktop completely
API Connection Issues
# Test Anthropic API
export ANTHROPIC_API_KEY=your-key
python -c "from anthropic import Anthropic; print(Anthropic().messages.create(model='claude-3-haiku-20240307', max_tokens=10, messages=[{'role':'user','content':'hi'}]))"
Common Issues
- "Command not found" - Ensure
produckai-mcpis in PATH - "Connection refused" - Check API keys are set
- "Import error" - Reinstall:
pip install --force-reinstall produckai-mcp-server
See docs/TROUBLESHOOTING.md for more.
🤝 Contributing
We welcome contributions! Here's how:
- Fork the repository
- Create a feature branch:
git checkout -b feature/your-feature - Commit your changes:
git commit -m "Add feature" - Push to your fork:
git push origin feature/your-feature - Open a Pull Request
See CONTRIBUTING.md for detailed guidelines.
Areas We Need Help:
- 📖 Documentation improvements
- 🐛 Bug fixes and testing
- ✨ New integration sources
- 🌍 Internationalization
- 🎨 UI/UX improvements
📊 Performance & Cost
Speed
- Feedback sync: ~1-2 seconds per item
- Clustering: ~1-2 minutes for 100 items
- PRD generation: ~10-15 seconds per PRD
Cost (AI APIs)
- Embeddings: ~$0.01 per 100 items (OpenAI)
- Clustering/Insights: ~$0.20 per 100 items (Claude Haiku)
- PRD Generation: ~$0.05-0.10 per PRD (Claude Sonnet)
- Monthly (100 PRDs): ~$5-10 total
📄 License
MIT License - see LICENSE for details.
🙏 Acknowledgments
- Built with MCP SDK
- Powered by Anthropic Claude
- Inspired by product teams everywhere
🔗 Links
- Issues - Report bugs
- Discussions - Ask questions
- Changelog - Release notes
- Security Policy - Report vulnerabilities
⭐ Star History
If you find this project useful, please star it! It helps others discover ProduckAI.
📧 Contact & Community
Get in Touch
- Creator: Rohit Saraf (rohitsaraff33@gmail.com)
- Issues: GitHub Issues - Bug reports, feature requests
- Discussions: GitHub Discussions - Questions, ideas, showcases
Vision & Community
This project was built for product managers, by product managers. The goal is to create a thriving open source community where builders enhance integrations, improve insight generation logic, and share learnings so the entire PM community benefits.
We especially welcome contributions in:
- 🔌 Integration enhancements - New data sources (Linear, Notion, Confluence, etc.)
- 🧠 Insight generation - Advanced clustering algorithms, sentiment analysis improvements
- 📊 Analytics & metrics - New VOC scoring dimensions, priority frameworks
- 📝 PRD templates - Industry-specific or company-specific variations
- 🌍 Localization - Multi-language support for global teams
Whether you're a PM improving your workflow or an engineer building better tools for PMs, your contributions help everyone in the product community. Let's build this together!
Built with ❤️ by Rohit Saraf and the product management community
Made with ❤️ by the ProduckAI community
