Aaasaasa CLI – Multi-Provider AI Tool

Aaasaasa CLI is a lightweight, fast, and extendable command-line interface
designed for working with multiple AI providers in a unified way.
It was built with the idea of combining local AI inference,
cluster-based AI scaling, and cloud AI APIs
into one seamless workflow.

🚀 Key Advantages

  • Unified interface for Ollama, VLLM clusters, and OpenAI
  • Supports multiple endpoints with advanced load balancing
  • Profiles system: switch between local, cluster, and cloud setups instantly
  • Lightweight & fast — written in Rust
  • Configurable with simple ~/.config/aaasaasa-cli/config.json
  • Ready for extension (custom providers, optimizers, balancing algorithms)

⚙️ Technologies

  • Rust for speed and safety
  • Tokio async runtime for concurrency
  • Reqwest for HTTP requests
  • Clap for CLI argument parsing
  • Serde for JSON configs

📂 Example Configuration

{
  "profiles": {
    "local-ollama": {
      "provider": "ollama",
      "endpoints": ["http://10.0.0.11:11434", "http://10.0.0.12:11434"],
      "balance": "least_conn",
      "model": "qwen2.5:7b-instruct-q5_K_M"
    },
    "cluster-vllm": {
      "provider": "vllm",
      "endpoints": ["http://10.0.1.21:8000", "http://10.0.1.22:8000"],
      "balance": "p2c",
      "model": "llama-3-70b-instruct"
    },
    "openai-fast": {
      "provider": "openai",
      "endpoint": "https://api.openai.com/v1/chat/completions",
      "balance": "round_robin",
      "model": "gpt-4o"
    }
  },
  "default_profile": "local-ollama"
}

💻 Usage Examples

# Start chat session with default profile
3a2a-cli chat "Hello 3a2a!"

# Use OpenAI profile
3a2a-cli -p openai-fast chat "Write me a poem."

# REPL interactive mode
3a2a-cli repl

# Show config path
3a2a-cli where

🌐 Real-World Use Cases

  • Run local inference for speed-sensitive tasks
  • Distribute workloads across multiple GPUs using cluster mode
  • Fallback to OpenAI API for high-accuracy tasks
  • Integrate into web applications and automation scripts

📈 Roadmap

  • AI-based auto-balancing between endpoints
  • Support for more providers (Anthropic, Mistral, etc.)
  • Cluster orchestration tools for large models
  • WebSocket & high-throughput pipelines

📜 License

This software is proprietary and licensed under the 3a2a™ EULA