Skip to content

v1.5.4

Latest
Latest

Choose a tag to compare

View all tags
@haiphucnguyen haiphucnguyen released this 19 Oct 04:22
· 19 commits to main since this release
v1.5.4
620ee7e
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Summary

This release focused on infrastructure modernization, migrating core components to newer technologies including SQLDelight for database management, cache4k for caching, and Kermit for logging. New features expand the platform's AI capabilities by integrating local Large Language Model (LLM) engines and enhancing voice processing with Whisper support. Additionally, several refactors improve system stability, standardize time handling, and update platform compatibility across various OS environments.

What's Changed

  • Database module now utilizes SQLDelight for schema definition and query handling, abstracting database operations for improved type safety and portability. (#742)
  • Configuration management was updated to migrate legacy Docker provider settings and standardize model request timeout parameters for API calls, ensuring consistent resource allocation. (#741, #746)
  • Caching implementation in the shared module was refactored to use the cache4k library, standardizing memory management and improving cache efficiency. (#744)
  • Logging infrastructure was updated to migrate the backend to Kermit logging and deprecate legacy event logging mechanisms, centralizing event capture and simplifying the logging pipeline. (#747, #751)
  • Project schema and Project structure were refined by removing redundant space fields, streamlining data storage and improving schema integrity. (#748)
  • MCP Instance handling was introduced to track and broadcast changes in the Model Control Plane (MCP) instance, enhancing system observability. (#745)
  • Time Handling was standardized across the application using Instant and Clock, resolving potential inconsistencies related to time zone management. (#752)
  • Voice Integration capabilities were expanded to include whispercpp for voice processing and model management, supporting bidirectional voice communication. (#756, #758, #774)
  • Local LLM functionality was added, providing integration support for external local Large Language Model engines, increasing deployment flexibility. (#762, #764)
  • Suggestions module received enhancements, implementing typing suggestion features for inline autocomplete and refining suggestion matching logic alongside usage history management. (#759, #761)
  • Core functionalities were enhanced by improving exception mapping and integrating detailed analytics reporting, enhancing system resilience and data capture. (#760)
  • Tool Handling was updated to manage the tool approval timeout state, preventing service blockages and ensuring robust tool lifecycle management. (#755)
  • UI/UX components received adjustments, including improvements to the chat interface, keymap handling, and integrating community forum links alongside improved error reporting. (#753, #773)
  • Networking/OpenRouter services were updated to load embedding models from a dedicated catalog, ensuring proper resource resolution. (#763)
  • Memory and Summarization logic was improved, optimizing memory retention and the efficiency of summarization processes. (#765)
  • Configuration updates implemented sparse config writing and updated dependency tracking, improving the efficiency of configuration persistence. (#766)
  • Whisper module received model download cancellation support, improving user experience during large model asset retrieval. (#767)
  • Analytics tracking was generalized and improved, enhancing the depth and reliability of usage metrics collection. (#768)
  • Text-to-Speech support was extended by adding the SystemTextToSpeech provider, broadening output modality options. (#770)
  • Platform Compatibility was improved by refining the macOS signing reliability and updating the Continuous Integration (CI) release workflow. (#771)
  • Local Storage management was refactored to use native JNA for the KeychainManager, ensuring platform-specific security best practices. (#772)
  • Voice Stream handling received improvements, adding support for live transcription and an associated UI toggle. (#775)
  • Recording flow was stabilized by fixing race conditions related to recording snapshots and improving cancellation handling. (#776)
  • Model Discovery and Internationalization (i18n) were improved for the LLM Browser component, expanding user access and language support. (#777)
  • UI/Navigation fixes included adjustments to hover height and logic to hide empty session headers, improving the visual consistency of the navigation sidebar. (#750)

Contributors

Full Changelog: v1.5.3...v1.5.4

Contributors

haiphucnguyen, jeonjw85, and 6 other contributors
Assets 13
Loading