Model-agnostic design philosophy
Emigo realizes three key benefits with the LiteLLM abstraction layer:
- Model Flexibility:Support for all models of the OpenRouter platform (e.g. Deepseek-chat/Claude, etc.), with users free to switch between them
- API compatibility:Harmonized handling of API specifications from different vendors, configuration files only need to modify the emigo-model parameter
- Cost Controllability:OpenRouter allows real-time comparison of the price/performance ratio of different models (e.g., Deepseek with ¥0.01/million tokens)
This architecture is already showing technical resilience in the beta phase. When testers find delays in response from one model, they can simply adjust the configuration to migrate to an alternative model with better performance, without modifying the code logic.
This answer comes from the articleEmigo: an assistant for complex programming tasks using AI in EmacsThe
May not be reproduced without permission:AI productivity tools " Emigo's Multi-Model Support Architecture Ensures Technical Foresight
