0.9.10

Added o1 model (untested) following its API access availability

0.9.9

• Incorporate API changes to use o1-mini (following ability to be able to test after access was relaxed around 20241120)
  • note that o1-mini and o1-preview are still beta models and provide limited control over features or capabilities. Those currently also do not support system role: if using prompt presets those might not work.

0.9.8

• Added "prompt presets" support (with example)
• Added "Clear Chat" button
• Extended Unraid template
• shift from requirements.txt to pyproject.toml and poetry

0.9.7

• Added gpt-4o-mini to the list of models
• deprecated older 32k models that were removed from the API page list

0.9.6

• Added method to disable vision for capable models
• added WebUI password protection using streamlit's secrets.toml method
• updated openai python to 1.35.6

0.9.5

• Added support for vision in capable models
• Added gpt-4-turbo models
• Deprecated some models in advance of 20240613
• Updated openai python package to 1.33.0
• Decoupled UI code to allow support for different frontends.

0.9.4

• Added support for gpt-4o
• updated openai python package to 1.29.0

0.9.3

• updated openai python package to 1.13.3
• Simplifying integration of new models and handling/presentation of their status (active, legacy, retired). The models.json file decouples development from model support
• Cleaner handling of max_tokens vs context window tokens

0.9.2

• updated openai python package: 1.8.0
• keep prompt history for a given session
• allow user to review/delete past prompts

0.9.1

• updated OpenAI to 1.3.3 (adds gpt-3.5-turbo-1106)
• make errors for environment variables shown on the UI
• cleanup syntax for independent models
• added an OAIWUI_USERNAME (not enabled by default)