Skip to main content

​
Overview

Finance Agent uses environment variables for configuration. Copy .env.example to .env and configure the values below.
Never commit .env to version control! It contains sensitive API keys and credentials.

​
Required Variables

​
AI/ML API Keys

​
OPENAI_API_KEY

OPENAI_API_KEY=sk-your-openai-api-key-here
Required: Yes Purpose: OpenAI API access for:
  • Embeddings generation (text-embedding-3-small)
  • Response generation (GPT-4 models)
  • Fallback LLM when Cerebras is unavailable
Get it: OpenAI API Keys

​
API_NINJAS_KEY

API_NINJAS_KEY=your-api-ninjas-key-here
Required: Yes Purpose: Access to earnings transcripts and financial data via API Ninjas Get it: API Ninjas

​
CEREBRAS_API_KEY

CEREBRAS_API_KEY=your-cerebras-api-key-here
Required: Yes (primary LLM) Purpose: Fast inference using Cerebras Cloud SDK
  • Primary LLM for RAG responses (Qwen-3-235B model)
  • Significantly faster than OpenAI for large-scale queries
Get it: Cerebras Cloud
Cerebras is the default LLM provider for optimal performance. Set RAG_LLM_PROVIDER=openai to use OpenAI instead.

​
Database Configuration

​
DATABASE_URL

DATABASE_URL=postgresql://username:password@localhost:5432/stratalens
Required: Yes Purpose: PostgreSQL connection string for:
  • User authentication and management
  • Chat history and conversations
  • Query analytics and logging
  • Application state
Format: postgresql://[username]:[password]@[host]:[port]/[database]
The database must have the pgvector extension installed. See Database Setup.

​
Authentication (Production)

​
CLERK_SECRET_KEY

CLERK_SECRET_KEY=sk_test_your-clerk-secret-key
Required: Yes (for production) Purpose: Clerk server-side authentication secret Get it: Clerk Dashboard → Your App → API Keys

​
CLERK_PUBLISHABLE_KEY

CLERK_PUBLISHABLE_KEY=pk_test_your-clerk-publishable-key
Required: Yes (for production) Purpose: Clerk client-side authentication key (safe to expose) Get it: Clerk Dashboard → Your App → API Keys

​
VITE_CLERK_PUBLISHABLE_KEY

VITE_CLERK_PUBLISHABLE_KEY=pk_test_your-clerk-publishable-key
Required: Yes (for frontend) Purpose: Same as CLERK_PUBLISHABLE_KEY, but with VITE_ prefix for Vite to expose to browser
This should have the same value as CLERK_PUBLISHABLE_KEY.

​
CLERK_WEBHOOK_SECRET

CLERK_WEBHOOK_SECRET=whsec_your-webhook-secret
Required: Optional Purpose: Verify Clerk webhook signatures for user events

​
Optional Variables

​
External Services

​
TAVILY_API_KEY

TAVILY_API_KEY=your-tavily-api-key-here
Required: Optional Purpose: Real-time web search for latest news and market developments Get it: Tavily

​
REDIS_URL

REDIS_URL=redis://localhost:6379
Required: Optional Purpose: Caching for WebSocket sessions and query results Default: redis://localhost:6379

​
LOGFIRE_TOKEN

LOGFIRE_TOKEN=your-logfire-token-here
Required: Optional Purpose: Application observability and monitoring via Logfire Get it: Logfire

​
Application Settings

​
ENVIRONMENT

ENVIRONMENT=development
Required: No Default: development Purpose: Environment detection for configuration Values:
  • development - Local development (increased connection limits, verbose logging)
  • production - Production deployment (stricter limits, optimized settings)

​
BASE_URL

BASE_URL=http://localhost:8000
Required: Yes Purpose: Server base URL for CORS and redirects Examples:
  • Development: http://localhost:8000
  • Production: https://www.stratalens.ai

​
PORT

PORT=8000
Required: No Default: 8000 Purpose: Server port

​
HOST

HOST=0.0.0.0
Required: No Default: 0.0.0.0 Purpose: Server host (0.0.0.0 allows external connections)

​
LOG_LEVEL

LOG_LEVEL=INFO
Required: No Default: INFO Purpose: Logging verbosity Values: DEBUG, INFO, WARNING, ERROR, CRITICAL

​
RAG System Configuration

​
RAG_LLM_PROVIDER

RAG_LLM_PROVIDER=cerebras
Required: No Default: cerebras Purpose: Choose LLM provider for RAG responses Values:
  • cerebras - Use Cerebras (fast, recommended)
  • openai - Use OpenAI GPT models
  • auto - Use Cerebras if CEREBRAS_API_KEY is set, else OpenAI

​
RAG_OPENAI_MAX_TOKENS

RAG_OPENAI_MAX_TOKENS=8000
Required: No Default: 8000 Purpose: Max completion tokens for OpenAI responses (lower = faster)

​
RAG_DEBUG_MODE

RAG_DEBUG_MODE=true
Required: No Default: false Purpose: Enable debug logging including EXPLAIN ANALYZE for database queries
Enable this during development to understand query performance and optimize database indexes.

​
RAG_MAX_ITERATIONS

RAG_MAX_ITERATIONS=3
Required: No Default: 3 Purpose: Maximum iterations for iterative self-improvement in RAG pipeline

​
SEC_MAX_ITERATIONS

SEC_MAX_ITERATIONS=5
Required: No Default: 5 Purpose: Maximum iterations for SEC 10-K filing queries (higher for complex financial analysis)

​
Feature Flags

​
AUTH_DISABLED

AUTH_DISABLED=true
Required: No Default: true Purpose: Bypass Clerk authentication (development only)
Never set this to true in production! Authentication should always be enabled for production deployments.

​
ENABLE_LOGIN

ENABLE_LOGIN=false
Required: No Default: false Purpose: Enable/disable login endpoint
This flag is deprecated. Authentication is now handled exclusively through Clerk.

​
ENABLE_SELF_SERVE_REGISTRATION

ENABLE_SELF_SERVE_REGISTRATION=false
Required: No Default: false Purpose: Enable/disable self-service user registration
This flag is deprecated. User registration is now handled through Clerk.

​
ENABLE_REGULAR_AUTH

ENABLE_REGULAR_AUTH=true
Required: No Default: true Purpose: Enable regular authentication flow

​
ENABLE_PREMIUM_ONBOARDING

ENABLE_PREMIUM_ONBOARDING=false
Required: No Default: false Purpose: Enable premium CSV-based user onboarding

​
Data Paths (Optional)

These variables are optional. The application uses sensible defaults if not specified.
# Override default data directories (optional)
DATA_DIR=./data
TRANSCRIPTS_DIR=./data/transcripts
EMBEDDINGS_DIR=./data/embeddings
10K_FILINGS_DIR=./data/10k_filings
DUCKDB_PATH=./data/duckdb/financial_data_new.duckdb

​
Legacy Authentication (Deprecated)

These variables are kept for backward compatibility during Clerk migration. They will be removed in a future release.
# JWT secret for legacy auth (deprecated)
JWT_SECRET_KEY=your-secret-key-here-change-in-production

​
Frontend-Specific Variables

​
VITE_API_BASE_URL

VITE_API_BASE_URL=
Required: No Default: Empty (same-origin requests) Purpose: API base URL for frontend requests
Leave this empty for same-origin requests (recommended for most setups). Only set it if your backend is on a separate domain.
Examples:
  • Same domain: Leave empty or set to empty string
  • Separate domain: https://api.stratalens.ai

​
Environment File Template

Here’s a complete .env template for local development: 
# ========================================
# AI/ML API Keys
# ========================================
OPENAI_API_KEY=sk-your-openai-api-key-here
API_NINJAS_KEY=your-api-ninjas-key-here
CEREBRAS_API_KEY=your-cerebras-api-key-here
TAVILY_API_KEY=your-tavily-api-key-here

# ========================================
# Database Configuration
# ========================================
DATABASE_URL=postgresql://username:password@localhost:5432/stratalens

# ========================================
# Application Settings
# ========================================
ENVIRONMENT=development
PORT=8000
HOST=0.0.0.0
BASE_URL=http://localhost:8000
LOG_LEVEL=INFO

# ========================================
# Clerk Authentication
# ========================================
CLERK_SECRET_KEY=sk_test_your-clerk-secret-key
CLERK_PUBLISHABLE_KEY=pk_test_your-clerk-publishable-key
VITE_CLERK_PUBLISHABLE_KEY=pk_test_your-clerk-publishable-key
CLERK_WEBHOOK_SECRET=whsec_your-webhook-secret

# ========================================
# Feature Flags
# ========================================
AUTH_DISABLED=true  # Development only!
RAG_DEBUG_MODE=true

# ========================================
# RAG System Configuration
# ========================================
RAG_LLM_PROVIDER=cerebras
RAG_OPENAI_MAX_TOKENS=8000
RAG_MAX_ITERATIONS=3
SEC_MAX_ITERATIONS=5

# ========================================
# External Services (Optional)
# ========================================
REDIS_URL=redis://localhost:6379
LOGFIRE_TOKEN=your-logfire-token-here

​
Next Steps

Build docs developers (and LLMs) love

Get started for freeTalk to us