Tabby

This patch release introduces a new configuration option to manage indexing for large installations. Users with over 20 repositories can now enable hourly sharded indexing using the TABBY_INDEX_REPO_IN_SHARD environment variable for better performance.

This is a patch release focusing primarily on stability improvements. The main change is the optimization of SQLite connection parameters to ensure a more stable operation. Users are advised to also review the full release notes for version 0.31.

This release introduces a significant new feature allowing Enterprise users to customize branding with their organization's name and logo. No bug fixes or general improvements were noted in this update.

This patch release focuses on stability and underlying performance improvements. Key updates include resolving database pool timeouts by upgrading sqlx and incorporating the latest llama.cpp changes. Users are also advised to review the full release notes for version 0.30.0.

This is a patch release focusing primarily on compatibility improvements. The main change is the downgrade of the CUDA base image to version 12.4.1. Users are advised to also review the full release notes for version 0.30.0.

This release introduces the ability to index GitLab Merge Requests as context, significantly improving suggestion relevance. Key improvements include leveraging the Answer Engine for better page generation quality and resolving UI flickering issues in chat responses. Note that the default Docker image now requires an NVIDIA GPU.

This release introduces powerful new RESTful APIs allowing users to programmatically ingest and remove custom documents for enhanced context. Key improvements include better error reporting in the Notification Inbox and the addition of Automatic Jobs status polling.

This release introduces the powerful "Convert to Page" feature, allowing users to transform Answer Engine responses into organized, shareable Pages. Additionally, support for Doc Query in the Chat Panel has been added, enabling the use of Developer Documentation as context.

This patch release focuses on stability and reliability, resolving critical issues that affected Windows users and offline server operations. Users should also review the full release notes for version 0.27 for broader context.

This release introduces powerful new capabilities for the Chat Panel, allowing users to execute shell commands and include uncommitted changes in context. Key fixes include resolving chat functionality issues with specific OpenAI reasoning models. Additionally, background job logging has been improved for better stability with large repositories.

This release introduces significant context capabilities by allowing the Answer Engine to access commit history. Users will also notice that their chat history is now visible on the Homepage and Side Panel. Additionally, documentation crawling has been improved when using llms-full.txt.

This is a patch release focusing on minor refinements rather than major new features or bug fixes. Users should review the full release notes for version 0.25.0 for comprehensive details. The primary visible change is an update to the Answer Engine UI.

This release introduces significant new capabilities for the Answer Engine, including exposing its thinking process and enabling repository file access. Users will also benefit from improved stability with automatic embedding retries and an enhanced user interface experience.

This release introduces LDAP Authentication Integration and adds notifications for failed background jobs. Several critical bugs were fixed, including issues with historical context in prompts, Jan AI integration, and model downloads from HuggingFace. Performance improvements include limiting history retention and optimizing GitHub PR diff indexing.

This patch release primarily addresses a critical bug that prevented the download of remote models following updates to the HuggingFace API. Users are advised to review the full release notes for version 0.23.0 for any other relevant changes.