🎨 LLM Slate

Web interface for chatting with 9+ AI providers in one place.

Chat with Claude, GPT-4, Grok, Gemini, and more from a single interface. Generate images with DALL-E or Aurora. Analyze images with Claude Vision or GPT-4 Vision. All with your own API keys, running locally.

Features

• Chat with 9 AI providers: Anthropic (Claude), OpenAI (GPT), xAI (Grok), Cohere, Mistral, Perplexity, Gemini, HuggingFace, Groq
• Image generation (DALL-E 3, Aurora)
• Image analysis (Claude Vision, GPT-4 Vision)
• Conversation history saved locally
• Password protected
• Clean, minimal interface

Quick Start

Installation

pip install llm-slate

Set Up API Keys

Create a .env file with your API keys:

# Add keys for the providers you want to use
ANTHROPIC_API_KEY=your_key_here
OPENAI_API_KEY=your_key_here
XAI_API_KEY=your_key_here
COHERE_API_KEY=your_key_here
MISTRAL_API_KEY=your_key_here
PERPLEXITY_API_KEY=your_key_here
GEMINI_API_KEY=your_key_here
HUGGINGFACE_API_KEY=your_key_here
GROQ_API_KEY=your_key_here

# Flask secret (generate with: python -c "import secrets; print(secrets.token_hex(32))")
FLASK_SECRET_KEY=your_secret_key_here

# Optional: Custom password (default: "friendship")
STUDIO_PASSWORD=your_password_here

# Optional: Custom port (default: 5413)
STUDIO_PORT=5413

Run

llm-slate

Then open http://localhost:5413 in your browser.

Usage

Chat

1. Select a provider from the dropdown
2. Choose a model
3. Start chatting

Your conversation history is saved and persists across sessions.

Image Generation

1. Go to the "Images" tab
2. Select provider (OpenAI or xAI)
3. Enter your prompt
4. Click generate

Image Analysis

1. Go to the "Vision" tab
2. Upload an image
3. Select provider (Anthropic or OpenAI)
4. Ask questions about the image

Configuration

Environment Variables

• STUDIO_PORT - Port to run on (default: 5413)
• STUDIO_PASSWORD - Login password (default: "friendship")
• STUDIO_BASE_PATH - Base path for reverse proxy (e.g., /studio)
• FLASK_SECRET_KEY - Flask session secret (required)

Provider API Keys

Only include keys for providers you want to use. Multi-Studio detects which providers are available based on your configured API keys.

Development

From Source

git clone https://github.com/lukeslp/llm-slate.git
cd llm-slate
pip install -e .

Project Structure

llm-slate/
├── app.py              # Main Flask application
├── config.py           # Configuration management
├── database.py         # Conversation storage
├── cache_manager.py    # Response caching
├── providers/          # Provider adapters
├── core/               # Core utilities
├── templates/          # HTML templates
├── static/             # CSS/JS assets
└── blueprints/         # Flask blueprints

Security

• All API keys are read from environment variables (never hardcoded)
• Conversations stored locally in SQLite database
• Password required to access interface
• No data sent to external services except your chosen AI providers

Supported Providers

Provider	Chat	Images	Vision
Anthropic (Claude)	✅	❌	✅
OpenAI (GPT)	✅	✅ (DALL-E)	✅
xAI (Grok)	✅	✅ (Aurora)	❌
Cohere	✅	❌	❌
Mistral	✅	❌	❌
Perplexity	✅	❌	❌
Gemini	✅	❌	❌
HuggingFace	✅	❌	❌
Groq	✅	❌	❌

Requirements

• Python 3.10+
• API keys for at least one provider
• Flask 3.0+

Author

Luke Steuber

• Website: lukesteuber.com
• Bluesky: @lukesteuber.com
• Email: luke@lukesteuber.com
• Support: Tip Jar

License

MIT License - see LICENSE for details.