Better init system

.gitignore

@@ -65,3 +65,6 @@ backends/advanced-backend/data/speaker_model_cache/
 *.bin
 *.sqlite3
 *checkpoints
+
+extras/openmemory-mcp/data/*
+.env.backup.*

CLAUDE.md

@@ -481,7 +481,9 @@ The sophisticated in-house memory implementation with full control and customiza
 - Individual fact storage (no JSON blobs)
 - Smart deduplication algorithms
 - Intelligent memory updates (ADD/UPDATE/DELETE decisions)
-- Direct Qdrant vector storage
+- **Semantic search** with relevance threshold filtering
+- **Memory count API** with total count tracking from native Qdrant
+- Direct Qdrant vector storage with accurate similarity scoring
 - Custom memory prompts and processing
 - No external dependencies
 
@@ -518,10 +520,12 @@ Thin client that delegates all memory processing to external OpenMemory MCP serv
 | **Processing** | Custom LLM extraction | Delegates to OpenMemory |
 | **Deduplication** | Custom algorithms | OpenMemory handles |
 | **Vector Storage** | Direct Qdrant | OpenMemory handles |
+| **Search Features** | Semantic search with threshold filtering | Semantic search with relevance scoring |
+| **Memory Count** | Native Qdrant count API | Varies by OpenMemory support |
 | **Dependencies** | Qdrant + MongoDB | External OpenMemory server |
 | **Customization** | Full control | Limited to OpenMemory features |
 | **Cross-client** | Friend-Lite only | Works with Claude Desktop, Cursor, etc |
-| **Web UI** | Friend-Lite WebUI | OpenMemory UI + Friend-Lite WebUI |
+| **Web UI** | Friend-Lite WebUI with advanced search | OpenMemory UI + Friend-Lite WebUI |
 | **Memory Format** | Individual facts | OpenMemory format |
 | **Setup Complexity** | Medium | High (external server required) |
 
@@ -640,6 +644,7 @@ Project includes `.cursor/rules/always-plan-first.mdc` requiring understanding b
 ### Memory & Conversation Debugging
 - **GET /api/admin/memories**: All memories across all users with debug stats (Admin only)
 - **GET /api/memories/unfiltered**: User's memories without filtering
+- **GET /api/memories/search**: Semantic memory search with relevance scoring
 - **GET /api/conversations**: User's conversations with transcripts
 - **GET /api/conversations/{audio_uuid}**: Specific conversation details
 

Docs/init-system.md

@@ -0,0 +1,207 @@
+# Friend-Lite Initialization System
+
+## Overview
+
+Friend-Lite uses a unified initialization system with clean separation of concerns:
+
+- **Configuration** (`init.py`) - Set up service configurations, API keys, and .env files
+- **Service Management** (`services.py`) - Start, stop, and manage running services
+
+The root orchestrator handles service selection and delegates configuration to individual service scripts. In general, setup scripts only configure and do not start services automatically. Exceptions: `extras/asr-services` and `extras/openmemory-mcp` are startup scripts. This prevents unnecessary resource usage and gives you control over when services actually run.
+
+## Architecture
+
+### Root Orchestrator
+- **Location**: `/init.py`
+- **Purpose**: Service selection and delegation only
+- **Does NOT**: Handle service-specific configuration or duplicate setup logic
+
+### Service Scripts
+- **Backend**: `backends/advanced/init.py` - Complete Python-based interactive setup
+- **Speaker Recognition**: `extras/speaker-recognition/setup.sh` - Simple bash setup
+- **ASR Services**: `extras/asr-services/setup.sh` - Service startup script  
+- **OpenMemory MCP**: `extras/openmemory-mcp/setup.sh` - External server startup
+
+## Usage
+
+### Orchestrated Setup (Recommended)
+Set up multiple services together with automatic URL coordination:
+
+```bash
+# From project root
+uv run --with-requirements setup-requirements.txt python init.py
+```
+
+The orchestrator will:
+1. Show service status and availability
+2. Let you select which services to configure
+3. Automatically pass service URLs between services
+4. Display next steps for starting services
+
+### Individual Service Setup
+Each service can be configured independently:
+
+```bash
+# Advanced Backend only
+cd backends/advanced
+uv run --with-requirements setup-requirements.txt python init.py
+
+# Speaker Recognition only  
+cd extras/speaker-recognition
+./setup.sh
+
+# ASR Services only
+cd extras/asr-services  
+./setup.sh
+
+# OpenMemory MCP only
+cd extras/openmemory-mcp
+./setup.sh
+```
+
+## Service Details
+
+### Advanced Backend
+- **Interactive setup** for authentication, LLM, transcription, and memory providers
+- **Accepts arguments**: `--speaker-service-url`, `--parakeet-asr-url`
+- **Generates**: Complete `.env` file with all required configuration
+- **Default ports**: Backend (8000), WebUI (5173)
+
+### Speaker Recognition  
+- **Prompts for**: Hugging Face token, compute mode (cpu/gpu)
+- **Service port**: 8085
+- **WebUI port**: 5173
+- **Requires**: HF_TOKEN for pyannote models
+
+### ASR Services
+- **Starts**: Parakeet ASR service via Docker Compose  
+- **Service port**: 8767
+- **Purpose**: Offline speech-to-text processing
+- **No configuration required**
+
+### OpenMemory MCP
+- **Starts**: External OpenMemory MCP server
+- **Service port**: 8765  
+- **WebUI**: Available at http://localhost:8765
+- **Purpose**: Cross-client memory compatibility
+
+## Automatic URL Coordination
+
+When using the orchestrated setup, service URLs are automatically configured:
+
+| Service Selected     | Backend Gets Configured With                                     |
+|----------------------|-------------------------------------------------------------------|
+| Speaker Recognition  | `SPEAKER_SERVICE_URL=http://host.docker.internal:8085`           |
+| ASR Services         | `PARAKEET_ASR_URL=http://host.docker.internal:8767`              |
+
+This eliminates the need to manually configure service URLs when running services on the same machine.
+Note (Linux): If `host.docker.internal` is unavailable, add `extra_hosts: - "host.docker.internal:host-gateway"` to the relevant services in `docker-compose.yml`.
+
+## Key Benefits
+
+✅ **No Unnecessary Building** - Services are only started when you explicitly request them  
+✅ **Resource Efficient** - Parakeet ASR won't start if you're using cloud transcription  
+✅ **Clean Separation** - Configuration vs service management are separate concerns  
+✅ **Unified Control** - Single command to start/stop all services  
+✅ **Selective Starting** - Choose which services to run based on your current needs
+
+## Service URLs
+
+### Default Service Endpoints
+- **Backend API**: http://localhost:8000
+- **Backend WebUI**: http://localhost:5173  
+- **Speaker Recognition**: http://localhost:8085
+- **Speaker Recognition WebUI**: http://localhost:5173
+- **Parakeet ASR**: http://localhost:8767
+- **OpenMemory MCP**: http://localhost:8765
+
+### Container-to-Container Communication
+Services use `host.docker.internal` for inter-container communication:
+- `http://host.docker.internal:8085` - Speaker Recognition
+- `http://host.docker.internal:8767` - Parakeet ASR  
+- `http://host.docker.internal:8765` - OpenMemory MCP
+
+## Service Management
+
+Friend-Lite now separates **configuration** from **service lifecycle management**:
+
+### Unified Service Management
+Use the `services.py` script for all service operations:
+
+```bash
+# Start all configured services
+uv run --with-requirements setup-requirements.txt python services.py start --all --build
+
+# Start specific services
+uv run --with-requirements setup-requirements.txt python services.py start backend speaker-recognition
+
+# Check service status
+uv run --with-requirements setup-requirements.txt python services.py status
+
+# Stop all services
+uv run --with-requirements setup-requirements.txt python services.py stop --all
+
+# Stop specific services  
+uv run --with-requirements setup-requirements.txt python services.py stop asr-services openmemory-mcp
+```
+
+### Manual Service Management
+You can also manage services individually:
+
+```bash
+# Advanced Backend
+cd backends/advanced && docker compose up --build -d
+
+# Speaker Recognition  
+cd extras/speaker-recognition && docker compose up --build -d
+
+# ASR Services (only if using offline transcription)
+cd extras/asr-services && docker compose up --build -d
+
+# OpenMemory MCP (only if using openmemory_mcp provider)
+cd extras/openmemory-mcp && docker compose up --build -d
+```
+
+## Configuration Files
+
+### Generated Files
+- `backends/advanced/.env` - Backend configuration with all services
+- `extras/speaker-recognition/.env` - Speaker service configuration
+- All services backup existing `.env` files automatically
+
+### Required Dependencies
+- **Root**: `setup-requirements.txt` (rich>=13.0.0)
+- **Backend**: `setup-requirements.txt` (rich>=13.0.0, pyyaml>=6.0.0)
+- **Extras**: No additional setup dependencies required
+
+## Troubleshooting
+
+### Common Issues
+- **Port conflicts**: Check if services are already running on default ports
+- **Permission errors**: Ensure scripts are executable (`chmod +x setup.sh`)
+- **Missing dependencies**: Install uv and ensure setup-requirements.txt dependencies available
+- **Service startup failures**: Check Docker is running and has sufficient resources
+
+### Service Health Checks
+```bash
+# Backend health
+curl http://localhost:8000/health
+
+# Speaker Recognition health  
+curl http://localhost:8085/health
+
+# ASR service health
+curl http://localhost:8767/health
+```
+
+### Logs and Debugging
+```bash
+# View service logs
+docker compose logs [service-name]
+
+# Backend logs
+cd backends/advanced && docker compose logs friend-backend
+
+# Speaker Recognition logs
+cd extras/speaker-recognition && docker compose logs speaker-service
+```

Docs/ports-and-access.md

@@ -0,0 +1,138 @@
+# Friend-Lite Port Configuration & User Journey
+
+## User Journey: Git Clone to Running Services
+
+### 1. Clone & Setup
+```bash
+git clone <repo>
+cd friend-lite
+
+# Configure all services
+uv run --with-requirements setup-requirements.txt python init.py
+
+# Start all configured services  
+uv run --with-requirements setup-requirements.txt python services.py start --all --build
+```
+
+### 2. Service Access Points
+
+## HTTP Mode (Default - No SSL Required)
+
+| Service | API Port | Web UI Port | Access URL |
+|---------|----------|-------------|------------|
+| **Advanced Backend** | 8000 | 5173 | http://localhost:8000 (API)<br>http://localhost:5173 (Dashboard) |
+| **Speaker Recognition** | 8085 | 5174 | http://localhost:8085 (API)<br>http://localhost:5174 (WebUI) |
+| **Parakeet ASR** | 8767 | - | http://localhost:8767 (API) |
+| **OpenMemory MCP** | 8765 | 8765 | http://localhost:8765 (API + WebUI) |
+
+**🌐 Main Dashboard**: http://localhost:5173  
+**🎤 Speaker Recognition**: http://localhost:5174  
+**❌ No microphone access** - browsers require HTTPS for microphone
+
+---
+
+## HTTPS Mode (For Microphone Access)
+
+| Service | HTTP Port | HTTPS Port | Access URL | Microphone Access |
+|---------|-----------|------------|------------|-------------------|
+| **Advanced Backend** | 80→443 | 443 | https://localhost/ (Main)<br>https://localhost/api/ (API) | ✅ Yes |
+| **Speaker Recognition** | 8081→8444 | 8444 | https://localhost:8444/ (Main)<br>https://localhost:8444/api/ (API) | ✅ Yes |
+
+**🌐 Main Dashboard**: https://localhost/ (Advanced Backend with SSL)  
+**🎤 Speaker Recognition**: https://localhost:8444/ (Speaker Recognition with SSL)  
+**✅ Full microphone access** - both services secured with SSL
+
+### Port Details (HTTPS Mode)
+- **Advanced Backend nginx**: Ports 80 (HTTP redirect) + 443 (HTTPS)
+- **Speaker Recognition nginx**: Ports 8081 (HTTP redirect) + 8444 (HTTPS)
+- **No port conflicts** - different port ranges for each service
+
+---
+
+## Why Two Modes?
+
+### HTTP Mode (Default)
+✅ **Simple setup** - No SSL certificates needed  
+✅ **Development friendly** - Quick start for testing  
+❌ **No microphone access** - Browsers require HTTPS for microphone
+
+### HTTPS Mode (Advanced)
+✅ **Microphone access** - Browsers allow mic access over HTTPS  
+✅ **Production ready** - Secure for real deployments  
+❌ **Complex setup** - Requires SSL certificate generation
+
+---
+
+## Configuration Files
+
+### Speaker Recognition Modes
+
+The speaker recognition service supports both modes via configuration:
+
+**HTTP Mode (.env)**:
+```bash
+REACT_UI_PORT=5174
+REACT_UI_HTTPS=false
+```
+
+**HTTPS Mode (.env)**:
+```bash
+REACT_UI_PORT=5175  # Internal HTTPS port
+REACT_UI_HTTPS=true
+# nginx provides external access on ports 8081 (HTTP) and 8444 (HTTPS)
+```
+
+---
+
+## Service Management Commands
+
+```bash
+# Check what's running
+uv run --with-requirements setup-requirements.txt python services.py status
+
+# Start all services
+uv run --with-requirements setup-requirements.txt python services.py start --all --build
+
+# Start only specific services
+uv run --with-requirements setup-requirements.txt python services.py start backend speaker-recognition
+
+# Stop all services
+uv run --with-requirements setup-requirements.txt python services.py stop --all
+```
+
+---
+
+## Microphone Access Requirements
+
+For **speaker recognition** and **live audio features** to work:
+
+1. **Local development**: Use HTTP mode, access via `http://localhost:5174`
+   - Some browsers allow localhost microphone access over HTTP
+
+2. **Production/Remote access**: Use HTTPS mode, access via `https://localhost:8444`
+   - All browsers require HTTPS for microphone access over network
+
+3. **Mixed setup**: Keep backend on HTTP, only enable HTTPS for speaker recognition when needed
+
+---
+
+## Port Conflict Resolution
+
+If you encounter port conflicts:
+
+1. **Check running services**: `uv run --with-requirements setup-requirements.txt python services.py status`
+2. **Stop conflicting services**: `uv run --with-requirements setup-requirements.txt python services.py stop --all`
+3. **Change ports in .env files** if needed
+4. **Restart services**: `uv run --with-requirements setup-requirements.txt python services.py start --all`
+
+---
+
+## Summary: Default User Experience
+
+After `git clone` and running init + services:
+
+🌐 **Main Application**: http://localhost:5173  
+🎤 **Speaker Recognition**: http://localhost:5174 (HTTP) or https://localhost:8444 (HTTPS)  
+🔧 **Backend API**: http://localhost:8000  
+📝 **ASR Service**: http://localhost:8767  
+🧠 **Memory Service**: http://localhost:8765