opensquilla

OpenSquilla is a token-efficient, microkernel AI agent built for CLI, Web UI, and various chat channels. Its core feature is an on-device model router (SquillaRouter) that dynamically dispatches each turn to the most cost-effective capable model. With persistent memory, a layered sandbox, built-in web search, and a pluggable provider layer supporting multiple LLMs, it maximizes intelligence density on the same budget.

• Dynamic Model Routing
• Token-Efficient Execution
• Persistent Memory & Sandbox
• Multi-Channel & Multi-LLM Integration