v2.3.1

Quality of Life

• LDAP authentication - due to frequent support requests, we've improved error messages and logging for LDAP authentication issues.

Bugfix

• Resolves an error occurring when executing php artisan route:cache, reading: Unable to prepare route [req/login] for serialization which was caused by two routes having the same name.

v2.3.0

What's New

• 🚨 IMPORTANT We made the development branch the new default branch. This does not affect your current setup, as main is still our release branch. But for new setups ensure you check out the main branch for production use (git clone -b main https://github.com/hawk-digital-environments/HAWKI.git). We updated our documentation to reflect this change.

Quality of Live

• Automatic CacheBuster for static assets. Instead of manually setting the asset version (js/css) in the templates, we now use a more robust hash generation, based on the HAWKI version and file modification time. If you extend HAWKI and want to hook into this logic? Take a look at the APP_CACHE_BUSTER environment variable.
• Dependency updates to the latest versions.
• Improve the error output in the ChainedAuthService
• (Docker) Ensure that PHP allows uploads greater than 2MB.

Bugfix

• Clean up the Dot Env section in the documentation to reflect the environment variables for the auth services and cache buster.
• Fix an issue in the LDAP config, where there were missing braces.
• Fixes an issue when using the external chat api without the "stream" property in the request.
• Fixes the app:token command to use the correct service.

Internal Changes

• Adds a release workflow and a changelog to the docs
• Inlines the documentation deployment pipeline into the HAWKI repository.

v2.2.0

v2.2.0

October 2025

What's New

• Custom Auth Services for your own needs; learn more in the documentation.

Quality of Life

• We created an extensive documentation about the Auth configuration here to help you understand what to configure where.
• LDAP environment variables. Over the last couple of weeks we heard about a lot of confusion because of the naming of the LDAP variables LDAP_BASE_DN, in comparison with the LDAP_SEARCH_DN and were asked why there was no LDAP_BIND_DN. We heared you, and restructured the variable names to make more sense in common LDAP termini; we hope this helps. fear not! As this is a minor update we provide full backward compatibility until the next major version.

Bugfix

• Auth Service cleanup - We did a lot of housekeeping tasks around the auth services (LDAP, Shibboleth and OIDC) to fix bugs but also allow a much cleaner flow.
• Updated docker service versions to the latest.

Deprecation

• The Shibboleth and OIDC routes (/req/login-shibboleth and /req/login-oidc) have been marked as deprecated and have been replaced with the shared route /req/login. Ensure to update your setup according to our documentation. fear not! As this is a minor update we provide full backward compatibility until the next major version.

v2.1.1

Changelog v2.1.۱

September 2025

This version introduces several new features, enhancements, and bug fixes.

Here is a summary of the latest changes in HAWKI:

What's New

1. Attachments
   
   Users can now add PDFs, Word documents, and images as attachments to chat messages. Files can be shared with other room members or sent to AI models as context, depending on model capabilities.

2. New File Storage System
   
   HAWKI now features a completely new file storage system. In addition to local storage, HAWKI supports SFTP, S3, and Nextcloud connectivity.

3. File Converter
   
   Since AI models cannot directly process document files, HAWKI now uses a new file conversion system to generate text-based context for AI models. You can either host the HAWKI Toolkit yourself or use the GWDG-hosted "Docling" service.

4. Backup System
   
   To ensure the security and integrity of your database data, HAWKI now includes a built-in backup system. You can activate scheduled backups and set the backup interval; HAWKI will automatically back up the database at regular intervals.

5. Scheduler
   As mentioned above HAWKI has now a schedule system as well. By running the scheduler service, HAWKI automatically takes care of Backups, checking model statuses, and garbage collection.

6. Announcements
   
   To keep users informed about the latest news, updates, or policy changes, you can now create Announcements. Users will see these as soon as they open HAWKI.

Quality of Life Improvements

We have received extensive feedback from our growing community and are committed to integrating your suggestions as quickly as possible. Here are some key improvements in HAWKI 2.1.0:

• Model Filtering
  Using toolset buttons on input fields, users can now view filtered lists of models based on features such as file upload, vision, and Google Search.

• Hidden Fields
  To enhance security, input fields containing passwords, data keys, etc., are now hidden. A custom hiding function also prevents browsers from misidentifying these fields as standard login fields.

• Customization
  As always, our goal is to make HAWKI as flexible as possible for every need. This version adds several new environment variables, enabling users to control various configuration options.

• Improved Broadcasting
  The new broadcasting logic allows chat rooms to instantly handle unlimited message sizes.

As HAWKI evolves, the codebase continues to grow. We remain committed to providing a well-structured foundation for our developer community. With this release, we have carefully restructured the codebase to facilitate further development. Although these changes may require adjustments for projects based on previous versions, they are intended to create a more sustainable and modular system for the future.

Bugfix:

This minor version includes fixes to some of the reported bugs in the main release (2.1.0):

• Fixed typo in profile reset function and route
• Fixed message edit function issue.
• Fixed the PostgreSQL migration problem.

v2.1.0

Changelog v2.1.0

September 2025

This version introduces several new features, enhancements, and bug fixes.

Here is a summary of the latest changes in HAWKI:

What's New

1. Attachments
   
   Users can now add PDFs, Word documents, and images as attachments to chat messages. Files can be shared with other room members or sent to AI models as context, depending on model capabilities.

2. New File Storage System
   
   HAWKI now features a completely new file storage system. In addition to local storage, HAWKI supports SFTP, S3, and Nextcloud connectivity.

3. File Converter
   
   Since AI models cannot directly process document files, HAWKI now uses a new file conversion system to generate text-based context for AI models. You can either host the HAWKI Toolkit yourself or use the GWDG-hosted "Docling" service.

4. Backup System
   
   To ensure the security and integrity of your database data, HAWKI now includes a built-in backup system. You can activate scheduled backups and set the backup interval; HAWKI will automatically back up the database at regular intervals.

5. Scheduler
   As mentioned above HAWKI has now a schedule system as well. By running the scheduler service, HAWKI automatically takes care of Backups, checking model statuses, and garbage collection.

6. Announcements
   
   To keep users informed about the latest news, updates, or policy changes, you can now create Announcements. Users will see these as soon as they open HAWKI.

Quality of Life Improvements

We have received extensive feedback from our growing community and are committed to integrating your suggestions as quickly as possible. Here are some key improvements in HAWKI 2.1.0:

• Model Filtering
  Using toolset buttons on input fields, users can now view filtered lists of models based on features such as file upload, vision, and Google Search.

• Hidden Fields
  To enhance security, input fields containing passwords, data keys, etc., are now hidden. A custom hiding function also prevents browsers from misidentifying these fields as standard login fields.

• Customization
  As always, our goal is to make HAWKI as flexible as possible for every need. This version adds several new environment variables, enabling users to control various configuration options.

• Improved Broadcasting
  The new broadcasting logic allows chat rooms to instantly handle unlimited message sizes.

As HAWKI evolves, the codebase continues to grow. We remain committed to providing a well-structured foundation for our developer community. With this release, we have carefully restructured the codebase to facilitate further development. Although these changes may require adjustments for projects based on previous versions, they are intended to create a more sustainable and modular system for the future.

v2.0.1

We are pleased to announce the release of HAWKI Version 2.0.1, featuring several important enhancements that improve functionality and user experience.

Key Updates and Features:

Optimized Model Integration:

Users now have the option to directly integrate locally hosted models into HAWKI, such as those managed through Ollama. This enhancement provides greater flexibility for privacy-sensitive scenarios and enables personalized setups. We extend our sincere gratitude to Sten Seegel from JLU Gießen for his exceptional work.

Internet Search Capability:

HAWKI can now perform internet searches! Through the integration of Google search functionality, the language model can access real-time web content while transparently displaying accurate source citations. We thank Sten once more for his instrumental support in implementing this feature.

Revised Onboarding Process:

We have improved the wording involved in creating the recovery key to enhance clarity and reduce confusion for new users. A special thank you goes to Raphael Fetzer from VCRP for providing valuable feedback on this matter.

Model Status at a Glance:

Given that some models may require additional time to start up or may be temporarily inaccessible, we have implemented a traffic light indicator before each model. This feature allows users to quickly ascertain a model's operational status at a glance.

HAWKI2 Docker Image:

An official Docker image for HAWKI2 is now available for users wishing to quickly and easily deploy or test the system locally. This feature offers a rapid setup, ideal for testing, demonstrations, or academic implementation. Special thanks to Martin Neundorfer for the Docker implementation and for his invaluable support and contribution to the project.

HAWKI CLI

The HAWKI Command Line Interface (CLI) provides a set of powerful commands for managing and interacting with your HAWKI installation. This tool simplifies common administrative tasks, system configuration, and maintenance operations through an intuitive command line interface.

Bug Fixes:

Numerous bugs have been fixed and minor improvements made, thanks to your continuous feedback!

---

We invite you to explore these new features and improvements. As always, we value your feedback as it helps us enhance the capabilities and user experience of HAWKI. For further details or to provide feedback, please visit our support page or contact us directly.

v2.0.0

Release Notes for HAWKI Version 2.0.0

We are excited to announce the release of HAWKI Version 2.0.0, a significant upgrade from the previous version. This milestone reflects substantial advancements and enhancements to both the functionality and user experience of the platform. The system is now built on the robust Laravel MVC framework, incorporating several key libraries to optimize user interaction and performance.

Key Updates and Features:

1. Enhanced User Experience and Stability:

   • Improvements have been made to the user interface to ensure a more intuitive and seamless interaction with the platform, along with increased overall stability.

2. Chat History Encryption:

   • All chat histories are now encrypted and securely stored in the database, ensuring user privacy and data protection in compliance with best security practices.

3. Introduction of Group Chat Module:

   • Users can now create and join group chat rooms, facilitating collaborative discussions and knowledge sharing using Generative AI tools and experiences.

4. AI-Driven Enhancements:

   • Integration of system AI for various functionalities, including:
     • Prompt Improvement: Suggestions to enhance the quality and effectiveness of user prompts.
     • Chat Name Generation: Automatic generation of suitable chat names to organize conversations better.
     • Chat Summary Generation: Provision of concise summaries when exporting chat conversations, aiding in efficient information retrieval and documentation.

These updates aim to foster collaboration, enhance productivity, and elevate the quality of interactions within the platform. We are committed to continuously improving HAWKI and welcome your feedback for future enhancements.

For more detailed information about this release or to provide feedback, please visit our support page or contact us directly.

v1.0.0-beta.1

Initial release of public beta version