Spegel's Multi-Model Support Architecture
Spegel achieves compatibility support for multiple AI models by integrating the litellm framework:
- Uses Google's Gemini 2.5 Flash Lite model by default
- Extended support for other major language models including GPT-4
- Allow the use of locally deployed large language models
This design provides ample options for users with different technical backgrounds. Developers can switch the underlying AI model through simple configuration file modifications without changing the main code.
Model selection affects the quality and speed of content conversion, and users can balance performance and effect according to specific scenarios. The default Gemini 2.5 Flash Lite can provide good conversion results in most scenarios, but for more professional text processing needs, more powerful models such as GPT-4 can bring more ideal conversion results.
This answer comes from the articleSpegel: using AI to transform web pages into an end-to-end browsing experienceThe
English 
简体中文 
日本語 
Deutsch 
Português do Brasil