What's Changed
- BREAKING: Now you can edit exist plugins by importing JAR file
- Added support for DeepSeek R1 models by detecting reasoning tokens
- Updated console.py by @KennySB-dev in #10
- Added support for OpenRouter referencer;
- Fixed historically incorrect terminology in README and code documentation, replacing "ChatGPT"/"GPT" with "LLM" to generalize model references;
- Changed the default model provider in
config.yaml
to OpenRouter.ai, with the default model updated todeepseek/deepseek-r1:free
; - Removed recommended tips and forced fallbacks for deprecated models (e.g.,
gpt-4-turbo
). We recommend using DeepSeek R1, OpenAI o1/o3-mini, or Gemini 2.0 Flash Thinking Exp 0121 instead; - Improved error handling for
APIConnectionError
,AuthenticationError
, and OpenRouter's Rate Limited, with clearer guidance prompts; - Added special handling for
o1-preview
(system prompts are unsupported in preview versions; strongly recommend usingo1
instead ofo1-preview
); - Deprecated the
disable_json_mode
parameter incore.askgpt
; - Logger now outputs to both log files and the console.
New Contributors
- @KennySB-dev made their first contribution in #10
Full Changelog: Beta-1.3...1.0.0