Enables containerized deployment of the complete conversation logging system with orchestrated services including API server, database, and caching layers
Stores conversation data, patterns, relationships, and insights across 8 specialized collections for comprehensive conversation management and analytics
Provides intelligent caching for conversation analysis results with 85%+ hit rate for fast response times and optimized performance
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Claude Conversation Loggersearch for recent conversations about authentication errors"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
π€ Claude Conversation Logger v3.1.0
π― Intelligent Conversation Management Platform - Advanced logging system with 4 Claude Code compatible agents, deep semantic analysis, automatic documentation, and 70% token optimization.
β 4 CLAUDE CODE AGENTS SYSTEM
π§ The Core Functionality
Claude Conversation Logger includes an optimized system of 4 Claude Code compatible agents that provides intelligent analysis, automatic documentation, and pattern discovery in technical conversations.
π The 4 Claude Code Agents
Agent | Primary Function | Use Cases |
π conversation-orchestrator-agent | Main coordinator making intelligent decisions | Multi-dimensional complex analysis, agent delegation |
π§ semantic-analyzer-agent | Deep semantic content analysis | Topics, entities, technical pattern extraction |
π pattern-discovery-agent | Historical pattern discovery | Identify recurring problems and solutions |
π auto-documentation-agent | Automatic documentation generation | Create structured problem-solution guides |
π Intelligent Capabilities
# π Intelligent semantic search
"authentication error" β Finds all authentication-related conversations
# π Contextual automatic documentation
Completed session β Automatically generates structured documentation
# π Intelligent relationship mapping
Current problem β Finds 5 similar conversations with solutions
# π Predictive pattern analysis
"API timeout" β Identifies 15 similar cases + most effective solutions
# π Multi-language support
Mixed ES/EN conversation β Detects patterns in both languagesβ‘ Key Benefits
β Token Optimization: 70% reduction vs manual analysis
β Instant Analysis: < 3 seconds for complete multi-agent analysis
β High Accuracy: 95%+ in pattern and state detection
β Multi-language Support: Spanish/English with extensible framework
β Intelligent Cache: 85%+ hit rate for fast responses
β Self-learning: Continuous improvement with usage
Related MCP server: TranscriptionTools MCP Server
π QUICK START - 3 STEPS
Step 1: Launch the System
# Clone and start
git clone https://github.com/LucianoRicardo737/claude-conversation-logger.git
cd claude-conversation-logger
# Launch with Docker (includes agents)
docker compose up -d --build
# Verify it's working
curl http://localhost:3003/healthStep 2: Configure Claude Code
# Copy MCP configuration
cp examples/claude-settings.json ~/.claude/settings.json
# Copy logging hook
cp examples/api-logger.py ~/.claude/hooks/
chmod +x ~/.claude/hooks/api-logger.pyStep 3: Use the Agents
# In Claude Code - search similar conversations
search_conversations({
query: "payment integration error",
days: 30,
includePatterns: true
})
# Intelligent analysis of current conversation
analyze_conversation_intelligence({
session_id: "current_session",
includeRelationships: true
})
# Automatic documentation
auto_document_session({
session_id: "completed_troubleshooting"
})π System ready! Agents are automatically analyzing all your conversations.
π CLAUDE CODE INTEGRATION (MCP)
5 Native Agent Tools
The system provides 5 native MCP tools for Claude Code:
MCP Tool | Responsible Agent | Functionality |
| semantic-analyzer-agent | Intelligent search with semantic analysis |
| conversation-orchestrator-agent | Recent activity with intelligent context |
| pattern-discovery-agent | Historical pattern analysis |
| auto-documentation-agent | Export with automatic documentation |
| conversation-orchestrator-agent | Complete multi-dimensional analysis |
Claude Code Configuration
~/.claude/settings.json
{
"mcp": {
"mcpServers": {
"conversation-logger": {
"command": "node",
"args": ["src/mcp-server.js"],
"cwd": "/path/to/claude-conversation-logger",
"env": {
"API_URL": "http://localhost:3003",
"API_KEY": "claude_api_secret_2024_change_me"
}
}
}
},
"hooks": {
"UserPromptSubmit": [{"hooks": [{"type": "command", "command": "python3 ~/.claude/hooks/api-logger.py"}]}],
"Stop": [{"hooks": [{"type": "command", "command": "python3 ~/.claude/hooks/api-logger.py"}]}]
}
}Claude Code Usage Examples
π Intelligent Search
// Search for similar problems with semantic analysis
search_conversations({
query: "React hydration mismatch SSR",
days: 60,
includePatterns: true,
minConfidence: 0.75
})
// Result: Related conversations + patterns + proven solutionsπ Pattern Analysis
// Identify recurring problems in project
analyze_conversation_patterns({
days: 30,
project: "my-api-service",
minFrequency: 3
})
// Result: Top issues + success rates + recommendationsπ Automatic Documentation
// Generate documentation from completed session
export_conversation({
session_id: "current_session",
format: "markdown",
includeCodeExamples: true
})
// Result: Structured markdown with problem + solution + codeπ§ Complete Multi-Agent Analysis
// Deep analysis with all agents
analyze_conversation_intelligence({
session_id: "complex_debugging_session",
includeSemanticAnalysis: true,
includeRelationships: true,
generateInsights: true
})
// Result: Complete analysis + insights + recommendationsπ οΈ AGENT REST API
5 Claude Code Endpoints
Analysis and Orchestration
# Complete multi-agent analysis
POST /api/agents/orchestrator
Content-Type: application/json
X-API-Key: claude_api_secret_2024_change_me
{
"type": "deep_analysis",
"data": {"session_id": "sess_123"},
"options": {
"includeSemanticAnalysis": true,
"generateInsights": true,
"maxTokenBudget": 150
}
}Pattern Discovery
# Find recurring patterns
GET /api/agents/patterns?days=30&minFrequency=3&project=api-service
# Response: Identified patterns + frequency + solutionsRelationship Mapping
# Search for related conversations
GET /api/agents/relationships/sess_123?minConfidence=0.7&maxResults=10
# Response: Similar conversations + relationship type + confidenceAutomatic Documentation
# Generate intelligent documentation
POST /api/agents/document
{
"session_id": "sess_123",
"options": {
"autoDetectPatterns": true,
"includeCodeExamples": true
}
}Main API Endpoints
Conversation Management
# Log conversation (used by hooks)
POST /api/conversations
# Search with semantic analysis
GET /api/conversations/search?q=authentication&days=30&semantic=true
# Export with automatic documentation
GET /api/conversations/{session_id}/export?format=markdown&enhanced=trueAnalytics and Metrics
# Project statistics
GET /api/projects/stats
# Agent metrics
GET /api/agents/metrics
# System health
GET /healthποΈ TECHNICAL ARCHITECTURE
Agent Architecture
graph TB
subgraph "π Claude Code Integration"
CC[Claude Code] -->|MCP Tools| MCP[MCP Server]
CC -->|Hooks| HOOK[Python Hooks]
end
subgraph "π€ Claude Code Agent System"
MCP --> CO[conversation-orchestrator-agent]
CO --> SA[semantic-analyzer-agent]
CO --> PD[pattern-discovery-agent]
CO --> AD[auto-documentation-agent]
end
subgraph "πΎ Data Layer"
SA --> MONGO[(MongoDB<br/>8 Collections)]
CO --> REDIS[(Redis<br/>Intelligent Cache)]
end
subgraph "π API Layer"
HOOK --> API[REST API Server]
API --> CO
end
style CO fill:#9c27b0,color:#fff
style SA fill:#2196f3,color:#fff
style MONGO fill:#4caf50,color:#fffSystem Components
Component | Technology | Port | Function |
π€ Agent System | Node.js 18+ | - | Intelligent conversation analysis |
π MCP Server | MCP SDK | stdio | Native Claude Code integration |
π REST API | Express.js | 3003 | Agent and management endpoints |
πΎ MongoDB | 7.0 | 27017 | 8 specialized collections |
β‘ Redis | 7.0 | 6379 | Intelligent agent cache |
π³ Docker | Compose | - | Monolithic orchestration |
Data Flow
sequenceDiagram
participant CC as Claude Code
participant MCP as MCP Server
participant CO as conversation-orchestrator-agent
participant AG as Agents (SA/PD/AD)
participant DB as MongoDB/Redis
CC->>MCP: search_conversations()
MCP->>CO: Process request
CO->>AG: Coordinate analysis
AG->>DB: Query data + cache
AG->>CO: Specialized results
CO->>MCP: Integrated response
MCP->>CC: Conversations + insightsβοΈ AGENT CONFIGURATION
42 Configuration Parameters
The agent system is fully configurable via Docker Compose:
π Language Configuration
# docker-compose.yml
environment:
# Primary languages
AGENT_PRIMARY_LANGUAGE: "es"
AGENT_SECONDARY_LANGUAGE: "en"
AGENT_MIXED_LANGUAGE_MODE: "true"
# Keywords in Spanish + English (JSON arrays)
AGENT_WRITE_KEYWORDS: '["documentar","guardar","document","save","create doc"]'
AGENT_READ_KEYWORDS: '["buscar","encontrar","similar","search","find","lookup"]'
AGENT_RESOLUTION_KEYWORDS: '["resuelto","funcionando","resolved","fixed","working"]'
AGENT_PROBLEM_KEYWORDS: '["error","problema","falla","bug","issue","crash"]'π― Performance Parameters
environment:
# Detection thresholds
AGENT_SIMILARITY_THRESHOLD: "0.75"
AGENT_CONFIDENCE_THRESHOLD: "0.80"
AGENT_MIN_PATTERN_FREQUENCY: "3"
# Token optimization
AGENT_MAX_TOKEN_BUDGET: "100"
AGENT_CACHE_TTL_SECONDS: "300"
# Feature flags
AGENT_ENABLE_SEMANTIC_ANALYSIS: "true"
AGENT_ENABLE_AUTO_DOCUMENTATION: "true"
AGENT_ENABLE_RELATIONSHIP_MAPPING: "true"
AGENT_ENABLE_PATTERN_PREDICTION: "true"8 Agent MongoDB Collections
Main Collections
// conversations - Base conversations
{
_id: ObjectId("..."),
session_id: "sess_123",
project: "api-service",
user_message: "Payment integration failing",
ai_response: "Let me help debug the payment flow...",
timestamp: ISODate("2025-08-25T10:00:00Z"),
metadata: {
resolved: true,
complexity: "intermediate",
topics: ["payment", "integration", "debugging"]
}
}
// conversation_patterns - Agent-detected patterns
{
pattern_id: "api_timeout_pattern",
title: "API Timeout Issues",
frequency: 23,
confidence: 0.87,
common_solution: "Increase timeout + add retry logic",
affected_projects: ["api-service", "payment-gateway"]
}
// conversation_relationships - Session connections
{
source_session: "sess_123",
target_session: "sess_456",
relationship_type: "similar_problem",
confidence_score: 0.89,
detected_by: "semantic-analyzer-agent"
}
// conversation_insights - Generated insights
{
insight_type: "recommendation",
priority: "high",
title: "Frequent Database Connection Issues",
recommendations: ["Add connection pooling", "Implement retry logic"]
}π§ INSTALLATION & DEPLOYMENT
Requirements
Docker 20.0+ with Docker Compose
Python 3.8+ (for hooks)
Claude Code installed and configured
4GB+ available RAM
Complete Installation
1. Clone and Setup
# Clone repository
git clone https://github.com/LucianoRicardo737/claude-conversation-logger.git
cd claude-conversation-logger
# Verify structure
ls -la # Should show: src/, config/, examples/, docker-compose.yml2. Docker Deployment
# Build and start complete system
docker compose up -d --build
# Verify services (should show 1 running container)
docker compose ps
# Verify system health
curl http://localhost:3003/health
# Expected: {"status":"healthy","services":{"api":"ok","mongodb":"ok","redis":"ok"}}3. Claude Code Configuration
# Create hooks directory if it doesn't exist
mkdir -p ~/.claude/hooks
# Copy logging hook
cp examples/api-logger.py ~/.claude/hooks/
chmod +x ~/.claude/hooks/api-logger.py
# Configure Claude Code settings
cp examples/claude-settings.json ~/.claude/settings.json
# Or merge with existing settings4. System Verification
# API test
curl -H "X-API-Key: claude_api_secret_2024_change_me" \
http://localhost:3003/api/conversations | jq .
# Agent test
curl -H "X-API-Key: claude_api_secret_2024_change_me" \
http://localhost:3003/api/agents/health
# Hook test (simulate)
python3 ~/.claude/hooks/api-logger.pyEnvironment Variables
Base Configuration
# Required
MONGODB_URI=mongodb://localhost:27017/conversations
REDIS_URL=redis://localhost:6379
API_KEY=your_secure_api_key_here
NODE_ENV=production
# Optional performance
API_MAX_CONNECTIONS=100
MONGODB_POOL_SIZE=20
REDIS_MESSAGE_LIMIT=10000Agent Configuration (42 variables)
# Languages and keywords
AGENT_PRIMARY_LANGUAGE=es
AGENT_MIXED_LANGUAGE_MODE=true
AGENT_WRITE_KEYWORDS='["documentar","document","save"]'
# Performance and thresholds
AGENT_MAX_TOKEN_BUDGET=100
AGENT_SIMILARITY_THRESHOLD=0.75
AGENT_CACHE_TTL_SECONDS=300
# Feature flags
AGENT_ENABLE_SEMANTIC_ANALYSIS=true
AGENT_ENABLE_AUTO_DOCUMENTATION=trueπ― PRACTICAL USE CASES
π Scenario 1: Recurring Debugging
// Problem: "Payments fail sporadically"
// In Claude Code, use MCP tool:
search_conversations({
query: "payment failed timeout integration",
days: 90,
includePatterns: true
})
// semantic-analyzer-agent + pattern-discovery-agent return:
// - 8 similar conversations found
// - Pattern identified: "Gateway timeout after 30s" (frequency: 23 times)
// - Proven solution: "Increase timeout to 60s + add retry" (success: 94%)
// - Related conversations: sess_456, sess_789, sess_012π Scenario 2: Automatic Documentation
// After solving a complex bug
// auto-documentation-agent generates contextual documentation:
export_conversation({
session_id: "debugging_session_456",
format: "markdown",
includeCodeExamples: true,
autoDetectPatterns: true
})
// System automatically generates:
/*
# Solution: Payment Gateway Timeout Issues
## Problem Identified
- Gateway timeout after 30 seconds
- Affects payments during peak hours
- Error: "ETIMEDOUT" in logs
## Investigation Performed
1. Nginx logs analysis
2. Timeout configuration review
3. Network latency monitoring
## Solution Implemented
```javascript
const paymentConfig = {
timeout: 60000, // Increased from 30s to 60s
retries: 3, // Added retry logic
backoff: 'exponential'
};Verification
β Tests passed: payment-integration.test.js
β Timeout reduced from 23 errors/day to 0
β Success rate: 99.2%
Tags
#payment #timeout #gateway #production-fix */
### **π Scenario 3: Project Analysis**
```javascript
// Analyze project health with pattern-discovery-agent
analyze_conversation_patterns({
project: "e-commerce-api",
days: 30,
minFrequency: 3,
includeSuccessRates: true
})
// System automatically identifies:
{
"top_issues": [
{
"pattern": "Database connection timeouts",
"frequency": 18,
"success_rate": 0.89,
"avg_resolution_time": "2.3 hours",
"recommended_action": "Implement connection pooling"
},
{
"pattern": "Redis cache misses",
"frequency": 12,
"success_rate": 0.92,
"avg_resolution_time": "45 minutes",
"recommended_action": "Review cache invalidation strategy"
}
],
"trending_topics": ["authentication", "api-rate-limiting", "database-performance"],
"recommendation": "Focus on database optimization - 60% of issues stem from DB layer"
}π Scenario 4: Intelligent Context Search
// Working on a new problem, search for similar context
// semantic-analyzer-agent finds intelligent connections:
search_conversations({
query: "React component not rendering after state update",
days: 60,
includeRelationships: true,
minConfidence: 0.7
})
// Result with relational analysis:
{
"direct_matches": [
{
"session_id": "sess_789",
"similarity": 0.94,
"relationship_type": "identical_problem",
"solution_confidence": 0.96,
"quick_solution": "Add useEffect dependency array"
}
],
"related_conversations": [
{
"session_id": "sess_234",
"similarity": 0.78,
"relationship_type": "similar_context",
"topic_overlap": ["React", "state management", "useEffect"]
}
],
"patterns_detected": {
"common_cause": "Missing useEffect dependencies",
"frequency": 15,
"success_rate": 0.93
}
}π§ Scenario 5: Complete Multi-Agent Analysis
// For complex conversations, activate all agents:
analyze_conversation_intelligence({
session_id: "complex_debugging_session",
includeSemanticAnalysis: true,
includeRelationships: true,
generateInsights: true,
maxTokenBudget: 200
})
// conversation-orchestrator-agent coordinates all agents:
{
"semantic_analysis": {
"topics": ["microservices", "docker", "kubernetes", "monitoring"],
"entities": ["Prometheus", "Grafana", "Helm charts"],
"complexity": "advanced",
"resolution_confidence": 0.91
},
"session_state": {
"status": "completed",
"quality_score": 0.87,
"documentation_ready": true
},
"relationships": [
{
"session_id": "sess_345",
"similarity": 0.82,
"type": "follow_up"
}
],
"patterns": {
"recurring_issue": "Kubernetes resource limits",
"frequency": 8,
"trend": "increasing"
},
"insights": [
{
"type": "recommendation",
"priority": "high",
"description": "Consider implementing HPA for dynamic scaling",
"confidence": 0.85
}
]
}π Complete Agent Documentation
For advanced usage and detailed configuration, consult the agent documentation:
π€ Claude Code Agent Files - Complete agent configurations in markdown format
π conversation-orchestrator-agent - Main orchestrator configuration
π§ semantic-analyzer-agent - Semantic analysis agent
π pattern-discovery-agent - Pattern discovery configuration
π auto-documentation-agent - Documentation generation agent
π Context System - Knowledge base and troubleshooting guides
π PROJECT STRUCTURE
claude-conversation-logger/
βββ π README.md # Main documentation
βββ π QUICK_START.md # Quick setup guide
βββ π³ docker-compose.yml # Complete orchestration
βββ π¦ package.json # Dependencies and scripts
βββ π§ config/ # Service configurations
β βββ supervisord.conf # Process management
β βββ mongodb.conf # MongoDB configuration
β βββ redis.conf # Redis configuration
βββ π src/ # Source code
β βββ server.js # Main API server
β βββ mcp-server.js # MCP server for Claude Code
β β
β βββ πΎ database/ # Data layer
β β βββ mongodb-agent-extension.js # MongoDB + agent collections
β β βββ redis.js # Intelligent cache
β β βββ agent-schemas.js # Agent schemas
β β
β βββ π§ services/ # Business services
β β βββ conversationService.js # Conversation management
β β βββ searchService.js # Semantic search
β β βββ exportService.js # Export with agents
β β
β βββ π οΈ utils/ # Utilities
β βββ recovery-manager.js # Data recovery
βββ π€ .claude/ # Claude Code Integration
β βββ agents/ # Agent definitions (markdown format)
β β βββ conversation-orchestrator-agent.md # Main orchestrator
β β βββ semantic-analyzer-agent.md # Semantic analysis
β β βββ pattern-discovery-agent.md # Pattern detection
β β βββ auto-documentation-agent.md # Documentation generation
β βββ context/ # Knowledge base and troubleshooting
βββ π‘ examples/ # Examples and configuration
β βββ claude-settings.json # Complete Claude Code config
β βββ api-logger.py # Logging hook
β βββ mcp-usage-examples.md # MCP usage examples
βββ π§ͺ tests/ # Test suite
βββ agents.test.js # Agent tests
βββ api.test.js # API tests
βββ integration.test.js # Integration testsπ METRICS & PERFORMANCE
π― Agent Metrics
Semantic Analysis: 95%+ accuracy in topic detection
State Detection: 90%+ accuracy in completed/active
Relationship Mapping: 85%+ accuracy in similarity
Token Optimization: 70% reduction vs manual analysis
Response Time: < 3 seconds complete analysis
β‘ System Performance
Startup Time: < 30 seconds complete container
API Response: < 100ms average
Cache Hit Rate: 85%+ on frequent queries
Memory Usage: ~768MB typical
Concurrent Users: 100+ supported
π Codebase Statistics
Lines of Code: 3,800+ (optimized agent system)
JavaScript Files: 15+ core files
Agent Files: 4 Claude Code compatible files
API Endpoints: 28+ endpoints (23 core + 5 agent tools)
MCP Tools: 5 native tools
MongoDB Collections: 8 specialized collections
π‘οΈ SECURITY & MAINTENANCE
π Security
API Key Authentication: Required for all endpoints
Helmet.js Security: Security headers and protections
Rate Limiting: 200 requests/15min in production
Configurable CORS: Cross-origin policies configurable
Data Encryption: Data encrypted at rest and in transit
π§ Troubleshooting
System won't start
# Check logs
docker compose logs -f
# Check resources
docker statsAgents not responding
# Agent health check
curl http://localhost:3003/api/agents/health
# Check configuration
curl http://localhost:3003/api/agents/configHook not working
# Manual hook test
python3 ~/.claude/hooks/api-logger.py
# Check permissions
chmod +x ~/.claude/hooks/api-logger.py
# Test API connectivity
curl -X POST http://localhost:3003/api/conversations \
-H "X-API-Key: claude_api_secret_2024_change_me" \
-H "Content-Type: application/json" \
-d '{"test": true}'π SUPPORT & CONTRIBUTION
π Get Help
π Technical Documentation: See Claude Code Agents
π Report Bugs: GitHub Issues
π‘ Request Features: GitHub Discussions
π€ Contribute
# Fork and clone
git clone https://github.com/your-username/claude-conversation-logger.git
# Create feature branch
git checkout -b feature/agent-improvements
# Develop and test
npm test
npm run test:agents
# Submit pull request
git push origin feature/agent-improvementsπ§ͺ Local Development
# Install dependencies
npm install
# Configure development environment
cp examples/claude-settings.json ~/.claude/settings.json
# Start in development mode
npm run dev
# Run agent tests
npm run test:agentsπ LICENSE & ATTRIBUTION
MIT License - See LICENSE for details.
Author: Luciano Emanuel Ricardo
Version: 3.1.0 - Claude Code Compatible Agent System
Repository: https://github.com/LucianoRicardo737/claude-conversation-logger
π EXECUTIVE SUMMARY
β
4 Claude Code Compatible Agents - Optimized multi-dimensional intelligent analysis
β
Native Claude Code Integration - 5 ready-to-use MCP tools
β
70% Token Optimization - Maximum efficiency in analysis
β
Multi-language Support - Spanish/English with extensible framework
β
Deep Semantic Analysis - True understanding of technical content
β
Automatic Documentation - Contextual guide generation
β
Pattern Discovery - Proactive identification of recurring problems
β
Relationship Mapping - Intelligent conversation connections
β
Intelligent Cache - 85%+ hit rate for instant responses
β
Complete REST API - 28+ endpoints including Claude Code agent tools
β
Docker Deployment - Production-ready monolithic system
β
42 Configurable Parameters - Complete customization via Docker Compose
π Ready for immediate deployment with intelligent agent system!
This server cannot be installed
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.