Skip to main content
LlamaFarm

LlamaFarm

Config-Based AI • Local First • Deploy Anywhere

The open-source framework for building, managing, and deploying AI applications with any model, on any cloud, anywhere. Take control of your AI infrastructure with simple YAML configurations.

📝

Configuration as Code

Define your entire AI infrastructure in simple YAML files. Version control your models, prompts, and deployment configurations. No more scattered scripts or manual setups.

🔄

True Model Portability

Switch between Llama, GPT-4, Claude, or any model with a single config change. Run locally for development, deploy to cloud for production. Your choice, your control.

🔒

Privacy by Default

Keep your data where it belongs - with you. Run models locally, process sensitive data on-premises, and only use cloud when you choose to. Complete data sovereignty.

Why LlamaFarm?

🏠 Local First

Run AI models locally on your hardware. No cloud dependency, no data leaving your infrastructure. Complete privacy and control.

🌍 Deploy Anywhere

One configuration, multiple destinations. Deploy to AWS, Azure, GCP, on-premises, or edge devices without changing your code.

🔧 Config-Based

Define your entire AI pipeline in simple YAML. No complex code, just declarative configurations that anyone can understand.

🤖 Any Model

Works with Llama, GPT, Claude, Mistral, and more. Switch models with a single config change. No vendor lock-in.

Simple as YAML

# llamafarm.yaml
models:
  - name: local-llama
    type: llama2
    path: ./models/llama-2-7b
    
  - name: cloud-gpt
    type: openai
    api_key: $OPENAI_KEY
    
pipeline:
  - input: user_query
  - model: local-llama
    fallback: cloud-gpt
  - output: response
    
deploy:
  targets:
    - local: true
    - aws: 
        region: us-east-1
    - edge:
        devices: ["rpi-cluster"]