Skip to content

Configuration

Aletheia is configured through environment variables. You can set these in a .env file or export them directly.

Database Configuration

FalkorDB (Default)

ALETHEIA_DATABASE_TYPE=falkordb
FALKORDB_HOST=localhost
FALKORDB_PORT=6379

Neo4j

ALETHEIA_DATABASE_TYPE=neo4j
NEO4J_URI=bolt://localhost:7687
NEO4J_USERNAME=neo4j
NEO4J_PASSWORD=password

LLM Configuration

OpenAI (Default)

OPENAI_API_KEY=sk-...

Dual LLM Setup

Aletheia supports separate models for different tasks:

# Reasoning model (schema inference, complex analysis)
ALETHEIA_REASONING_MODEL=o3-mini

# Fast model (extraction, evaluation)
ALETHEIA_FAST_MODEL=gpt-4o-mini

Local Embeddings

To use local embeddings instead of OpenAI:

ALETHEIA_USE_LOCAL_EMBEDDINGS=true

Evaluation Configuration

Search Configuration

# Default search limit
ALETHEIA_SEARCH_LIMIT=15

# BFS traversal depth
ALETHEIA_BFS_MAX_DEPTH=3

Grounding Configuration

# Default grounding mode
ALETHEIA_GROUNDING_MODE=strict  # strict, lenient, off

Example .env File

# Database
ALETHEIA_DATABASE_TYPE=falkordb
FALKORDB_HOST=localhost
FALKORDB_PORT=6379

# LLM
OPENAI_API_KEY=sk-your-key-here
ALETHEIA_REASONING_MODEL=o3-mini
ALETHEIA_FAST_MODEL=gpt-4o-mini

# Evaluation
ALETHEIA_SEARCH_LIMIT=15
ALETHEIA_GROUNDING_MODE=strict

Docker Compose Example

For local development with FalkorDB:

version: '3.8'
services:
  falkordb:
    image: falkordb/falkordb
    ports:
      - "6379:6379"
    volumes:
      - falkordb_data:/data

volumes:
  falkordb_data:

Learn More