Installation

Installation

Install LM Deluge using pip (Python 3.10+):

python -m pip install -U lm-deluge

Optional extras:

• pip install plyvel if you want LevelDB-backed local caching
• pip install pdf2image pillow if you plan to turn PDFs into images via Image.from_pdf

API Keys

LM Deluge reads API keys from environment variables so the client can contact each provider directly. Load them at process startup (for example with python-dotenv) and pass the values down to your workers, CLI scripts, or notebook kernels.

Required Environment Variables

Depending on which providers you plan to use, set the appropriate API keys:

# OpenAI
OPENAI_API_KEY=sk-...

# Anthropic
ANTHROPIC_API_KEY=sk-ant-...

# Google AI Studio (Gemini)
GEMINI_API_KEY=...

# Cohere
COHERE_API_KEY=...

# OpenRouter (any model prefixed with openrouter:)
OPENROUTER_API_KEY=...

# Meta, Groq, DeepSeek, etc. use provider-specific keys defined in src/lm_deluge/models

# AWS Bedrock (for Amazon-hosted Anthropic and Meta models)
AWS_ACCESS_KEY_ID=...
AWS_SECRET_ACCESS_KEY=...

.env File

Create a .env file in your project root:

.env

OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GEMINI_API_KEY=...

Load the file inside your application:

import dotenv

dotenv.load_dotenv()

Verification

Verify your installation by running a simple test:

import dotenv
from lm_deluge import LLMClient

dotenv.load_dotenv()

client = LLMClient("gpt-4.1-mini")
responses = client.process_prompts_sync(["Say hello!"])
print(responses[0].completion)

If you see a response from the model, you’re all set!

Next Steps

Head over to the Quick Start guide to learn how to use LM Deluge effectively.