OpenClaw Talk Analyzer

AI-Powered Business Conversation Analysis

Analyze business conversations and generate actionable insights - An intelligent tool for extracting summaries, strategies, and key points from meeting transcripts, sales calls, and customer interactions.

Transform raw conversation data into structured business intelligence with AI-powered analysis.

Features

• Multi-Source Analysis - Process conversations from various sources (audio transcripts, chat logs, meeting recordings)
• AI-Powered Insights - Leverage Claude, GPT, or local LLMs for intelligent analysis
• Sentiment Analysis - Detect emotional tone and engagement levels
• Action Items Extraction - Automatically identify tasks, decisions, and commitments
• Speaker Profiling - Analyze individual speaking patterns and contribution metrics
• Strategy Recommendations - Generate data-driven suggestions for follow-up actions
• Export Reports - Generate formatted reports in JSON, Markdown, or PDF

Supported Analysis Types

| Type | Description | Status | |------|-------------|--------| | Meeting Summary | Extract key points, decisions, and action items | ✅ | | Sales Call Analysis | Identify objections, opportunities, and next steps | ✅ | | Customer Support | Analyze sentiment, resolution quality, and satisfaction | ✅ | | Negotiation Strategy | Evaluate positions, concessions, and leverage points | ✅ | | Team Dynamics | Assess participation, collaboration patterns | ⏳ | | Sentiment Tracking | Monitor emotional trends across conversations | ✅ |

Quick Start

1. Installation

# Clone the repository
git clone https://github.com/yourusername/openclaw-talk-analyzer.git
cd openclaw-talk-analyzer
npm install

# Or install globally
npm install -g openclaw-talk-analyzer

2. Configure API Keys

# Copy configuration template
cp .env.example .env

# Edit .env with your AI provider credentials
nano .env

Add your API keys:

# Claude (recommended)
ANTHROPIC_API_KEY=your-api-key

# Or OpenAI
OPENAI_API_KEY=your-api-key

# Or use local LLM
LOCAL_LLM_ENDPOINT=http://localhost:11434

3. Analyze Your First Conversation

# Analyze a transcript file
openclaw-talk analyze \
  --input meeting-transcript.txt \
  --type meeting \
  --output report.json

# Analyze with specific focus
openclaw-talk analyze \
  --input sales-call.txt \
  --type sales \
  --focus "objections,pricing,next-steps"

# Generate detailed report
openclaw-talk analyze \
  --input conversation.txt \
  --type negotiation \
  --output report.md \
  --format markdown \
  --detailed

Usage Guide

Basic Analysis

openclaw-talk analyze \
  --input <file-path> \
  --type <analysis-type> \
  --output <output-file>

Analysis Types:

• meeting - General meeting summary
• sales - Sales call analysis
• support - Customer support interaction
• negotiation - Negotiation strategy analysis
• general - General conversation analysis

Advanced Options

openclaw-talk analyze \
  --input conversation.txt \
  --type sales \
  --focus "objections,pricing,timeline" \
  --sentiment \
  --speakers "John,Sarah,Mike" \
  --language en \
  --model claude-sonnet-4 \
  --output detailed-report.json

Options:

• --focus - Specific aspects to analyze (comma-separated)
• --sentiment - Enable sentiment analysis
• --speakers - List of speaker names for attribution
• --language - Input language (default: auto-detect)
• --model - AI model to use (claude-sonnet-4, gpt-4, local)
• --format - Output format (json, markdown, pdf)
• --detailed - Include detailed analysis and quotes

Batch Analysis

Process multiple conversations:

# Analyze all files in a directory
openclaw-talk batch \
  --input ./transcripts/*.txt \
  --type meeting \
  --output ./reports/

# Use configuration file
openclaw-talk batch --config batch-config.json

Example batch-config.json:

{
  "files": [
    {
      "input": "meeting-2024-01-15.txt",
      "type": "meeting",
      "speakers": ["Alice", "Bob", "Charlie"]
    },
    {
      "input": "sales-call-acme.txt",
      "type": "sales",
      "focus": ["objections", "budget", "timeline"]
    }
  ],
  "options": {
    "sentiment": true,
    "format": "markdown",
    "model": "claude-sonnet-4"
  }
}

Compare Conversations

# Compare multiple conversations to identify patterns
openclaw-talk compare \
  --inputs "call1.txt,call2.txt,call3.txt" \
  --type sales \
  --output comparison-report.md

Output Examples

Meeting Summary (JSON)

{
  "type": "meeting",
  "date": "2024-03-10",
  "participants": ["Alice", "Bob", "Charlie"],
  "duration": "45 minutes",
  "summary": "Team discussed Q2 product roadmap priorities...",
  "key_points": [
    "Launch mobile app by end of Q2",
    "Hire 2 additional engineers",
    "Increase marketing budget by 20%"
  ],
  "action_items": [
    {
      "task": "Finalize mobile app wireframes",
      "owner": "Alice",
      "deadline": "2024-03-20"
    },
    {
      "task": "Draft job descriptions for engineering roles",
      "owner": "Bob",
      "deadline": "2024-03-15"
    }
  ],
  "decisions": [
    "Approved mobile app development budget: $150k"
  ],
  "sentiment": {
    "overall": "positive",
    "participants": {
      "Alice": "optimistic",
      "Bob": "neutral",
      "Charlie": "enthusiastic"
    }
  },
  "next_meeting": "2024-03-24"
}

Sales Call Analysis (Markdown)

# Sales Call Analysis - Acme Corp

**Date:** 2024-03-10
**Duration:** 32 minutes
**Participants:** Sarah (Sales), John (Prospect - CTO)

## Summary
Strong initial interest in the enterprise plan. Prospect expressed concerns about integration complexity but is excited about automation features.

## Key Opportunities
- High budget authority ($200k+)
- Clear pain points with current solution
- Timeline pressure (Q2 launch needed)

## Objections Raised
1. **Integration complexity** - Concerned about API limitations
   - *Response needed:* Technical demo with their existing stack
2. **Pricing** - Requested volume discount
   - *Recommendation:* Offer 15% discount for annual commitment

## Next Steps
- [ ] Schedule technical demo (by March 15)
- [ ] Send pricing proposal with discount (by March 12)
- [ ] Connect prospect with existing customer in similar industry

## Strategy Recommendation
**Priority: HIGH** - Strong fit, high urgency. Focus on quick technical validation and competitive pricing to close by end of month.

Project Structure

talk-analyzer/
├── src/
│   ├── core/
│   │   ├── types.ts              # TypeScript type definitions
│   │   ├── analyzer.ts           # Main analysis orchestrator
│   │   └── processor.ts          # Text preprocessing
│   ├── engines/
│   │   ├── base.ts               # Base analyzer engine
│   │   ├── claude.ts             # Claude API integration
│   │   ├── openai.ts             # OpenAI API integration
│   │   └── local.ts              # Local LLM integration
│   ├── analyzers/
│   │   ├── sentiment.ts          # Sentiment analysis
│   │   ├── speakers.ts           # Speaker profiling
│   │   ├── actions.ts            # Action items extraction
│   │   └── strategy.ts           # Strategy recommendations
│   ├── exporters/
│   │   ├── json.ts               # JSON export
│   │   ├── markdown.ts           # Markdown export
│   │   └── pdf.ts                # PDF export
│   └── cli/
│       └── index.ts              # CLI entry point
├── examples/
│   ├── sample-meeting.txt
│   ├── sample-sales-call.txt
│   └── batch-config.json
├── config/
│   └── prompts/                  # AI prompt templates
├── tests/
└── README.md

Configuration

Environment Variables (.env)

# AI Provider (choose one)
ANTHROPIC_API_KEY=your-claude-api-key
OPENAI_API_KEY=your-openai-api-key
LOCAL_LLM_ENDPOINT=http://localhost:11434

# Default Settings
DEFAULT_MODEL=claude-sonnet-4
DEFAULT_FORMAT=json
ENABLE_SENTIMENT=true
MAX_TOKENS=4096

# Output
OUTPUT_DIR=./reports
LOG_LEVEL=info

Use Cases

1. Sales Teams

• Analyze call recordings to identify best practices
• Track objection patterns across prospects
• Generate follow-up strategies automatically

2. Product Managers

• Extract feature requests from customer interviews
• Identify common pain points
• Prioritize development based on conversation insights

3. Customer Success

• Monitor support conversation quality
• Track customer sentiment over time
• Identify at-risk accounts early

4. Executives

• Get concise summaries of important meetings
• Track team decision-making patterns
• Identify strategic opportunities

Advanced Features

Custom Analysis Prompts

Create custom analysis templates:

// config/prompts/custom-analysis.ts
export const customPrompt = {
  name: "partnership-evaluation",
  prompt: `Analyze this conversation for:
    1. Strategic alignment
    2. Resource requirements
    3. Risk factors
    4. Mutual benefits
    Provide actionable recommendations.`
};

API Integration

Use as a library in your application:

import { TalkAnalyzer } from 'openclaw-talk-analyzer';

const analyzer = new TalkAnalyzer({
  apiKey: process.env.ANTHROPIC_API_KEY,
  model: 'claude-sonnet-4'
});

const result = await analyzer.analyze({
  text: conversationText,
  type: 'sales',
  options: {
    sentiment: true,
    speakers: ['Alice', 'Bob']
  }
});

console.log(result.summary);
console.log(result.action_items);

Roadmap

• [ ] Real-time audio transcription integration
• [ ] Multi-language support (currently English only)
• [ ] Web dashboard for visualization
• [ ] Team collaboration features
• [ ] Integration with CRM systems (Salesforce, HubSpot)
• [ ] Slack/Teams bot integration

Privacy & Security

• All analysis is processed securely
• No conversation data is stored on external servers (unless using cloud AI APIs)
• Local LLM option for sensitive conversations
• GDPR and CCPA compliant

Contributing

Contributions welcome! Please read CONTRIBUTING.md for guidelines.

License

MIT License - see LICENSE for details

Support

• Documentation: Full docs
• Issues: GitHub Issues
• Discord: Join our community

Transform conversations into actionable business intelligence.