v1.3.2

Others

• Update gpt-4o model
• Set MAX_MODEL_CONTEXT in service/.env file

v1.3.1

Features

• Auto-save slider values for advanced settings

Fixes

• Style of reordering list
• Language adaptability for title generation

v1.3.0

Features

• Generate summary title for the current conversation
• Auto-generate summary title on starting a conversation
• Move model selection to the sidebar

Performance

• Separate chat storage into state and data to speed up UI responsiveness
• Reduce startup loading time by delayed permission check

Fixes

• Fix the stop responding button to the bottom
• Scroll the conversation list to the top when creating a new chat

v1.2.1

Features

• Replace model gpt-3.5-turbo with gpt-4o-mini

Others

• [backend] Add error message for invalid model

v1.2.0

Features

• Add a search box to search chat list by title
• Allow dragging to reorder the chat list
• Hover on the chat list showing the full title

Fixes

• Unexpected message editing state on context change
• Recover animation of setting page show-up

v1.1.1

Fixes

• Auto-focus editing box on edit
• Bad behaviour when there is no existing conversation

Performance

• Reduce lagging when typing prompts
• Remove startup message loading check

v1.1.0

Features

• Conversation-wise draft prompt and context state [@Gnefil]
• [backend] Customise running port in .env file

Fixes

• Use a static url path without chat id
• Avoid frontend error when server is unavailable
• Bad page positioning when typing on ios
• Missing large file caching from PWA service worker
• Reset title editing state on reload
• [backend] API endpoint binding

Refactor

• Chat store and chat process with semantic APIs
• File and directory structure

v1.0.1

Features

• Support conversation-level context switching [@Gnefil]

Fixes

• Payload size limit with long chat history
• Unexpected regeneration without valid user prompt

v1.0.0

Features

Frontend

• Model selection (gpt-4 / gpt-3.5)
• In-place message editing [@Gnefil]
• Full session history control

Backend

• Built with official openai-node API
• Multi-user authentication and profiles
• Model-wise token usage statistics logging for users