AI Model Support in MoltBot

MoltBot is designed to work with multiple AI models, giving you the flexibility to choose the best option for your needs. Whether you prefer cloud-based models or want to run everything locally, MoltBot has you covered.

Supported AI Models

Anthropic Claude

Claude is MoltBot's default AI backend, known for:

• Excellent reasoning capabilities
• Strong safety features
• Long context window
• Natural conversation style

OpenAI GPT

Full support for GPT models including:

• GPT-4 and GPT-4 Turbo
• GPT-3.5 Turbo
• Custom fine-tuned models

Local Models

Run AI completely offline with:

• LLaMA and LLaMA 2
• Mistral
• Phi-2
• Any GGUF format model

Choosing the Right Model

For Best Performance

Claude and GPT-4 offer the most capable responses, ideal for:

• Complex reasoning tasks
• Code generation
• Creative writing

For Privacy

Local models keep everything on your machine:

• No data sent to external servers
• Works offline
• Complete privacy

For Cost Efficiency

GPT-3.5 and smaller local models are great for:

• High-volume tasks
• Simple queries
• Budget-conscious usage

Configuration

Set your preferred model in the config:

# Use Claude (default)
moltbot config set model claude

# Use GPT-4
moltbot config set model gpt-4

# Use a local model
moltbot config set model local/llama-2-7b

API Keys

For cloud models, you'll need API keys:

moltbot config set anthropic_key YOUR_KEY
moltbot config set openai_key YOUR_KEY

Model Switching

MoltBot can automatically switch between models based on:

• Task complexity
• API availability
• Cost constraints

Configure fallback options:

moltbot config set fallback_model gpt-3.5-turbo

Performance Tips

1. Use local models for simple, repetitive tasks
2. Reserve powerful cloud models for complex reasoning
3. Set up model routing for optimal cost/performance balance