#270 · Primary category: AI Agents & Automation

opensquilla

agent ai ai-agents deep-learning foundation-models llm mcp memory openclaw python skills

OpenSquilla — Token-Efficient AI Agent with same budget, higher intelligence density

Project last updated:08/30/26

GitHub Stars

6.8K

Forks

534

Contributors

48

License

Apache-2.0

Why we included this project

OpenSquilla is an agent that tries to stretch your LLM budget by routing each turn to the cheapest model that can handle it. A small on-device router picks the right provider from your configured list, with a fallback to direct single-model routing if the local runtime isn't fully set up. The same turn loop drives the web UI, CLI, and chat channels, so tool dispatch, retries, and decision logs stay consistent across surfaces, and persistent memory, a layered sandbox, built-in web search, and local embeddings come along for the ride. Its provider layer covers OpenRouter, OpenAI, Anthropic, Ollama, DeepSeek, Gemini, Qwen, and more than 20 other endpoints without schema changes, which helps if you juggle multiple backends. It also ships as packaged desktop installers in addition to a terminal tool, so you can pick the interface you prefer.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category