All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Added Anthropic Sonnet v2 model
- Use newest Google Gemini models
- Default to gpt-4o-2024-08-06 model
- Added gpt-4o-mini model
- Set document retrieval limit to 15 and threshold to 0.35
- Use gpt-4o-mini for notion batch processing
- Fix citations in consecutive messages
- Fix citations
- Fix citation links
- Fix citations outside \cite{} blocks.
- Set retrieval threshold to 0.30
- Update to service version 8.0.0
- Fix multiline on mobile and web
- Enabled multiline prompts
- Default to Claude Sonnet model
- Upgrade to Claude 3.5 Sonnet
- Performance and bug fixes
- Increased the maximum output length to 2048 tokens
- Updated Google Gemini Pro models
- Added GPT 4o model
- Fix single line code blocks
- Added Database selection
- Added Notion API management
- Fix Claude 3 Opus model
- Added notion support
- Fix bold text in chat
- Update generating animation
- Fixed inline markdown code blocks
- Small bug fixes
- Fixed some latex bugs
- Fixed spelling
- Added support for code blocks in markdown
- Updated OpenAI GPT-4 Turbo model
- Added Gemini 1.5 model
- Remove Mistral AI models
- Added Claude 3 Haiku
- Removed Claude 2
- Added Claude Opus
- Fixed latex rendering in square brackets
- Added Claude 3 Sonnet
- Added Mistral model
- Added crashlytics
- Fix strange line breaking with latex formulas
- Fixed balance sheet
- Remove firebase crashlytics due to missing support for web
- Added Firebase Crashlytics
- Updated settings
- Added document chunk search
- The reference text can now be viewed
- Added support for website scraping
- Added support for document references in text
- Fixed a bug that prevented static document selection
- Changed xushengs/flutter_markdown_latex to pzierahn/flutter_markdown_latex
- Fix bug that caused involuntary document selection
- Added support for latex formulas
- Update models
- Fix a bug that prevented the showing of error messages in the chat
- Added support for AWS Bedrock models
- Update protos
- Added a warning when no documents are uploaded to a collection
- Added references to thread history
- Prevent follow-up questions when content is generated
- Update settings page
- Added support follow-up questions
- Fall back to gpt-4-1106-preview model
- Update to newest GPT-4 model
- Update default parameters for fragment search
- Update protos
- Added a Delete-Button for chat messages
- Added a Remove-Button on error messages
- Simplify model selection
- Fix page number in sources
- Fixed document page issue
- Added Google Gemini Pro model
- Changed protobuf
- Fix file upload for web
- Use Google Cloud instead of Supabase
- Remove GPT-4
- Rename Temperature to Creativity
- Rename Limit to Input Pages
- Remove Threshold Slider
- Remove max Tokens Slider
- Increased the maximum number of pages to 100
- Fix upload button in collection view
- Added gpt-4-1106-preview model to the model list
- Show funding request errors in the UI
- Added a waring message if no documents are uploaded to a collection
- Added a payments overview in user settings