Documentation Index Fetch the complete documentation index at: https://docs.callintel.io/llms.txt
Use this file to discover all available pages before exploring further.
Environment Variables All configuration is managed through environment variables. Create a .env file in the project root:
# Required: Environment type Environment=local # or production Queue_Environment=local # AI Provider APIs GEMINI_API_KEY=your_gemini_key OPENAI_API_KEY=your_openai_key # Hume AI (Emotional Intelligence) HUME_CONFIG_ID=your_config_id HUME_API_KEY=your_hume_key HUME_SECRET_KEY=your_hume_secret # Telephony Providers TWILIO_ACCOUNT_ID=your_twilio_sid TWILIO_AUTH_TOKEN=your_twilio_token # Alternative: Plivo PLIVO_AUTH_ID=your_plivo_id PLIVO_AUTH_TOKEN=your_plivo_token # Database SUPABASE_URL=https://xxx.supabase.co SUPABASE_KEY=your_supabase_key # MCP Service URLs (Docker) WHATSAPP_MCP_SERVER_URL=http://whatsapp_mcp:8102 TELE_MCP_SERVER_URL=http://tele_mcp:8100 GMAIL_MCP_SERVER_URL=http://gmail_mcp:8101 # Server Configuration PORT=8010 SERVER_DOMAIN=localhost:8010 SECURE_HEADERS=false CORS_ORIGINS=* # Logging LOG_LEVEL=INFO # Optional: LiveKit (for real-time communication) LIVEKIT_URL=ws://livekit:7880 LIVEKIT_API_KEY=your_livekit_key LIVEKIT_API_SECRET=your_livekit_secret
Configuration Files Local Configuration File: configure/local_config.py
import os from dotenv import load_dotenv load_dotenv() # Logging LOG_LEVEL = os.getenv( "LOG_LEVEL" , "INFO" ) # API Configuration API_HOST = os.getenv( "SERVER_DOMAIN" , "localhost:8010" ) API_PORT = int (os.getenv( "PORT" , 8010 )) # Database SUPABASE_URL = os.getenv( "SUPABASE_URL" ) SUPABASE_KEY = os.getenv( "SUPABASE_KEY" ) # AI Providers GEMINI_API_KEY = os.getenv( "GEMINI_API_KEY" ) OPENAI_API_KEY = os.getenv( "OPENAI_API_KEY" ) # Telephony TWILIO_ACCOUNT_ID = os.getenv( "TWILIO_ACCOUNT_ID" ) TWILIO_AUTH_TOKEN = os.getenv( "TWILIO_AUTH_TOKEN" ) # MCP Services MCP_SERVICES = { "telegram" : os.getenv( "TELE_MCP_SERVER_URL" , "http://localhost:8100" ), "gmail" : os.getenv( "GMAIL_MCP_SERVER_URL" , "http://localhost:8101" ), "whatsapp" : os.getenv( "WHATSAPP_MCP_SERVER_URL" , "http://localhost:8102" ), }
Production Configuration File: configure/production_config.py
Similar structure but with production-specific settings:
Disabled API documentation (docs_url=None)
Stricter CORS configuration
Enhanced logging for monitoring
Database connection pooling
AI Provider Configuration Google Gemini from bridges.gemini_bridge_base import GeminiBridgeBase agent = GeminiBridgeBase( api_key = "your_gemini_key" , model = "gemini-2.0-flash-exp" , # Latest model system_prompt = "Your custom system prompt" , temperature = 0.7 , max_tokens = 1024 )
Required Settings:
GEMINI_API_KEY: From Google Cloud ConsoleModel selection: gemini-2.0-flash-exp, gemini-pro, etc.
OpenAI from bridges.open_ai import OpenAIBridge agent = OpenAIBridge( api_key = "your_openai_key" , model = "gpt-4" , system_prompt = "Your custom system prompt" , temperature = 0.7 )
Required Settings:
OPENAI_API_KEY: From OpenAI API dashboardModel: gpt-4, gpt-3.5-turbo, gpt-4-turbo, etc.
Hume AI (Emotional Intelligence) from bridges.hume_bridge import HumeBridge agent = HumeBridge( config_id = "your_config_id" , api_key = "your_hume_key" , secret_key = "your_hume_secret" )
Required Settings:
HUME_CONFIG_ID: Configuration ID from Hume AIHUME_API_KEY: Public API keyHUME_SECRET_KEY: Secret key for authentication
Telephony Configuration Twilio from twilio.rest import Client client = Client( account_sid = os.getenv( "TWILIO_ACCOUNT_ID" ), auth_token = os.getenv( "TWILIO_AUTH_TOKEN" ) )
Setup Steps:
Create Twilio account at twilio.com
Get Account SID and Auth Token
Purchase phone number or verify test number
Configure webhook URL for incoming calls
Plivo from plivo import RestClient client = RestClient( auth_id = os.getenv( "PLIVO_AUTH_ID" ), auth_token = os.getenv( "PLIVO_AUTH_TOKEN" ) )
Setup Steps:
Create Plivo account
Get Auth ID and Token
Purchase DID number
Set up message URL
Database Configuration Supabase from Service.Supabase import get_supabase_client supabase = get_supabase_client() # Tables are auto-created on first run # Tables include: # - organizations # - agents # - calls # - contact_forms # - scheduled_calls # - batch_calls
Setup Steps:
Create Supabase project
Copy Project URL and API Key
Set SUPABASE_URL and SUPABASE_KEY
Server Configuration Port Configuration PORT=8010 SERVER_DOMAIN=localhost:8010
CORS Configuration # In app.py app.add_middleware( CORSMiddleware, allow_origins = os.getenv( "CORS_ORIGINS" , "*" ).split( "," ), allow_credentials = True , allow_methods = [ "*" ], allow_headers = [ "*" ], )
API Documentation Local Environment:
Swagger UI: /docs
ReDoc: /redoc
OpenAPI: /openapi.json
Production Environment:
Documentation disabled (docs_url=None)
Use external tools like Postman for API testing
Logging Configuration Log Levels LOG_LEVEL=INFO # DEBUG, INFO, WARNING, ERROR, CRITICAL
Log Output from utils.logger import logger logger.debug( "Debug message" ) logger.info( "Info message" ) logger.warning( "Warning message" ) logger.error( "Error message" ) logger.critical( "Critical message" )
Logs are written to:
Console (stdout)
File: logs/app.log (if configured)
MCP Service Configuration MCP (Model Context Protocol) services handle external integrations:
Telegram MCP TELE_MCP_SERVER_URL=http://tele_mcp:8100
Configuration file: integrations/mcp_tele/config.json
Gmail MCP GMAIL_MCP_SERVER_URL=http://gmail_mcp:8101
OAuth credentials: integrations/mcp_gmail/credentials.json
WhatsApp MCP WHATSAPP_MCP_SERVER_URL=http://whatsapp_mcp:8102
Database configuration: integrations/mcp_whatsapp/database.py
Agent Configuration System Prompts DEFAULT_PROMPT_TEMPLATE = """ Assistant Design: " {initial_message} " Core Identity & Purpose: Primary Mission: To be a helpful, human-like voice assistant Interaction Style: - Professional and friendly - Natural conversation flow - Problem-solving focused """
Voice Configuration voice_config = { "provider" : "twilio" , # or "elevenlabs", "google" "voice_id" : "en-US" , "speed" : 1.0 , "pitch" : 1.0 , }
Security Configuration API Keys Best practices:
Use environment variables, never commit keys
Rotate keys regularly
Use different keys for dev/prod
Store in .env file (add to .gitignore)
CORS CORS_ORIGINS = [ "https://yourdomain.com" , "https://app.yourdomain.com" , ]
Authentication from utils.auth import verify_supabase_token @app.get ( "/api/protected" ) async def protected_route ( user : AuthenticatedUser = Depends(verify_supabase_token)): return { "user" : user.id}
Worker Configuration # Gunicorn (if using production server) WORKERS=4 WORKER_CLASS=uvicorn.workers.UvicornWorker WORKER_CONNECTIONS=1000
Database Connection Pooling SUPABASE_POOL_SIZE = 10 SUPABASE_MAX_OVERFLOW = 20
Cache Configuration REDIS_URL = redis: // localhost: 6379 CACHE_TTL = 3600
Integration Configuration Shopify SHOPIFY_API_KEY=your_shopify_key SHOPIFY_API_SECRET=your_shopify_secret SHOPIFY_STORE_URL=your-store.myshopify.com
Payment Provider STRIPE_API_KEY=your_stripe_key STRIPE_WEBHOOK_SECRET=your_webhook_secret
Environment-Specific Configs Development Environment=local LOG_LEVEL=DEBUG CORS_ORIGINS=* DEBUG=true
Staging Environment=production LOG_LEVEL=INFO CORS_ORIGINS=https://staging.yourdomain.com DEBUG=false
Production Environment=production LOG_LEVEL=WARNING CORS_ORIGINS=https://yourdomain.com DEBUG=false SECURE_HEADERS=true
Validation Verify your configuration:
# Check if all required environment variables are set python -c "from configure.local_config import *; print('Configuration loaded successfully')" # Test API connectivity curl http://localhost:8010/health # Test database connection python -c "from Service.Supabase import get_supabase_client; print(get_supabase_client())"
Common Issues Missing API Keys # Generate .env from example cp .env.example .env # Check required variables grep "TODO\|CHANGE_ME" .env
Connection Failures # Test Supabase curl https://your-project.supabase.co/health # Test Twilio python -c "from twilio.rest import Client; Client('sid', 'token').api.accounts.list()"
Next Steps
Quick Start Start building with the configured platform.
