- Create simplified BPE tokenizer based on tiktoken core logic
- Use real production metrics: 3,144 tokens/s prompt, 121 tokens/s generation
- Add accurate processing time estimates for translation ETA
- Add HTTP error checking for 404 endpoint issues
- Ensure models are loaded before translation attempts
- Add response validation with clear error messages
- Fix floating button DOM timing in preview windows
- Remove modal/dialog DOM from translation preview content
- Fix 404 errors by properly checking model registry for endpoints
- Enable floating button in preview windows with partial initialization
- Reverted the broken openTranslateModal function to a working state.
- Re-implemented the logic to correctly inject the <base> tag for relative paths and to ensure the translation notice is displayed correctly.
- Ran biome to format the code and ensure there are no syntax errors.
- The placeholder restoration logic now processes placeholders in reverse order, ensuring that nested content (like links within HTML tags) is restored correctly.
- The translation preview now includes the full HTML document, ensuring that stylesheets, scripts, and other resources are loaded correctly.
- A <base> tag is injected into the preview to handle relative resource paths.
- Removed a script flag that was incorrectly disabling JavaScript in the translated page, which fixes the issue with dialog boxes not appearing.
- Add createTranslateFeature() function for standalone translate button
- Include 'translate' in custom interface feature switch
- Add translate-only demo section with live example
- Update documentation to show translate in available features
- Translation now works as data-features='translate' or part of 'full'
- Update table of contents and section numbering
- Update to 19 languages actually supported by Hermes 3 model
- Match Russell's placeholder naming: __CODE_X__, __URI_X__, __HTML_X__
- Use exact preservation technique from translate_content.py
- Add Bengali, Urdu, Indonesian, Swahili, Marathi, Telugu
- Include Traditional Chinese (zh-tw) support
- Match Team Hermes implementation approach
- Add translation.js with 20+ language support
- Implement code block, URL, and HTML tag preservation during translation
- Add translation modal with page translation and custom text modes
- Integrate translation button in all interfaces
- Add translation demo section with live examples
- Support Russell's approach of preserving technical content during translation
- Enable translation of current page content or custom text
- Fix modal conversation persistence between sessions
- Fix system messages appearing in DOM by filtering display
- Add configurable system message with setSystemMessageAppend()
- Reorganize demo.html with table of contents and code-example pattern
- Add comprehensive documentation for system message customization
- Make all code blocks open by default in demo page
- Show 'loading...' text in floating button while modal opens
- Display modal immediately with loading placeholders
- Load models and Hermes intro asynchronously after modal appears
- Prevent blocking UI during slow network operations
- Fix model dropdown to return correct model names instead of full unique IDs
- Replace problematic DOM traversal in TTS buttons with proper event handlers
- Prevent model dropdown creation on pages without chat interfaces
- Restore full-featured Hermes modal from original implementation
- Add comprehensive modal with model/voice selection, action buttons, and chat interface
- Include conversation history management and message deletion functionality
new file: src/config.js
new file: src/content.js
new file: src/file-upload.js
new file: src/models.js
new file: src/page-reader.js
new file: src/storage.js
new file: src/tts.js
new file: src/ui.js
modified: uncloseai.js
new file: uncloseai.js.orig
Exposed , , , , , , , , , and functions to the global object. This resolves issues when these functions are called directly from HTML attributes, particularly in .
This commit introduces a function to generate unique localStorage keys based on the current page's URL. This key is now used for storing and retrieving conversation history, ensuring that each page maintains its own isolated chat thread. The Hermes introduction already uses a content-based hash for per-page caching, and this change aligns the conversation history with that behavior. Additionally, clearing the conversation history now only affects the current page and correctly repopulates the Hermes introduction for that page.
- Add .uncloseai class support with data-features configuration
- Create shared utility functions for buttons, chat messages, TTS controls, voice selects, and modal headers
- Refactor class-based feature creators to use shared utilities
- Add comprehensive documentation for class-based integration in index.html
- Add live examples of class-based usage in demo.html
- Maintain backward compatibility with existing floating button and API functions
