Enthusiast v1.4.2

Enthusiast v1.4.2

Enthusiast v1.4.2 is out! This release focuses on improving file handling, stability, and documentation consistency across the project - along with a few UX and API upgrades to make things smoother for contributors and users alike.

Highlights

• File Handling & Agent Improvements

  • Fixed the FILE_UPLOAD flag in OCR and catalog enrichment agents.
  • Improved the file upload UX for a more intuitive experience.
  • Fixed catalog enrichment agent prompts for better accuracy.
  • Ensured Medusa-based tools handle default products without categories or properties.

• API & Configuration Updates

  • Included Medusa API variables in tool configuration for easier setup.
  • Added region parameter to the order tool for more flexible deployments.
  • Moved parsers to server-side code for better organization.

New Contributors

• @magdalena-k-up made their first contribution in #214

Full Changelog: v1.4.1...v1.4.2

v1.4.1

Enthusiast v1.4.1

This patch release focuses on stability, minor fixes, and improving package consistency. It includes updates to the Azure OpenAI plugin and dependency bumps to ensure smoother runtime behavior and developer workflows.

Highlights

Maintenance & Dependency Updates

• Updated language model provider packages versions to keep dependencies up to date and compatible.
• Renamed the Azure OpenAI plugin directory for consistency and clarity.

Full Changelog: v1.4.0…v1.4.1

v1.4.0

Enthusiast v1.4.0

Enthusiast v1.4.0 is here! This version brings new features and possibilities by enabling users to upload files and use it as a context in conversation.

Highlights

• File Handling

  • Created API endpoints and UI for uploading files to a conversation.
  • Added common file parsers and automatic processing of uploaded files.
  • Added ability to enable passing files content to agents.

• Documentation & UI

  • Integrated Nextra for improved docs.
  • Reorganized documentation structure and added integrations section.
  • Refactored chat and added new message types.
  • Added example agents use cases plugins.

New Contributors

• @fpdrozd made their first contribution in #200

Full Changelog: Compare v1.3.0…v1.4.0

v1.3.0

Enthusiast v1.3.0

We’re excited to release Enthusiast v1.3.0! This version introduces major improvements to agent management, expanded model support, and a more unified UI experience. We’ve also added new configuration and plugin options to make it easier to customize and extend Enthusiast.

Highlights

• Expanded Model & Plugin Support
  • Added GPT-4.1 to prioritized models.
  • Added Mistral plugin.
  • Added Azure OpenAI connector.
• Agent Management Improvements
  • Agents section in the admin UI for managing and configuring agents.
  • Custom agents can now be parameterized, allowing for reuse and easier experimentation.
  • Validate Agent configurations and set default agents for new datasets.
• UI & Usability Upgrades
  • Unified UI across sections for a consistent look and feel.
  • Redirect to new chat page after login for a smoother experience.
  • Edit dataset configurations directly from the interface.
  • Trigger automatic sync tasks after adding a source.
• Developer Experience & Infrastructure
  • Improved Swagger documentation coverage for all API endpoints.
  • New docs for Agent memory, prompts, configuration, architecture, tools, and injectors.
  • Improved handling of JSON input arguments and plugin configuration.
  • Updated frontend and docs dependencies.

New Contributors

• @PWiercioch made their first contribution in #143

Full Changelog: v1.2.3...v1.3.0

v1.2.3

Enthusiast v1.2.3

This release introduces a series of enhancements aimed at improving the internal architecture and developer experience. Key improvements include better interface design, structured build steps for memory and prompt handling, and a more robust default behavior for the ReAct agent.

Highlights

• Improved Agent Configuration
  • Added build steps for LLM callback handlers and prompt construction to streamline agent setup.
  • Memory instance construction is now managed via the injector’s build step for better modularity.
• Improved Templates
  • ReAct agent now includes a default get_answer implementation for more consistent behavior out of the box.
• Interface & Dependency Improvements
  • Refined injector and common interfaces to make code navigation easier.

Full Changelog: v1.2.2...v1.2.3

v1.2.2

Enthusiast v1.2.2

This patch release focuses on stability, configuration reliability, and keeping dependencies up to date.

Highlights

• Tools now use a structured input by default, allowing them to accept multiple parameters.
• Added template implementation for ReAct agents.

Full Changelog: v1.2.1...v1.2.2

v1.2.1

Enthusiast v1.2.1

This patch release focuses on stability, configuration reliability, and keeping dependencies up to date. It includes several small but important fixes to ensure smoother runtime behavior and developer workflows.

Highlights

• Improved Agent Configuration:
  • Added merge support in the Agent registry to handle layered configurations.
  • Default builder fallback makes it easier to create a custom agent.
  • LLM configuration is now included in the default settings for more robust defaults.
• Embedding Generation Fixes
• Maintenance & Dependency Updates:
  • Lockfiles for plugins regenerated for consistency.
  • Moved enthusiast server to use poetry for simplified setup and better dependency management

Full Changelog: v1.2.0...v1.2.1

v1.2.0

Enthusiast v1.2.0

We’re excited to release Enthusiast v1.2.0! This update brings a range of usability improvements, enhanced developer tooling, and important infrastructure updates to support smoother testing and deployment workflows.

Highlights

• Better Source Management UX: The source management screen has been refined for a more intuitive user experience.
• Agent Selection in Chat: Users can now choose which agent to use directly in the chat interface.
• Chat UI Enhancements: Improved scroll behavior and a new “Start New Chat” button for better usability.
• Improved Developer Experience: Predefined agent memory and configurations to help you get started faster.
• Internal Improvements:
  • Error handling: Agent messages are now saved properly if there's an execution error.
  • Type definitions cleanup: Minor fixes including type definitions and validation cleanup.
  • Pydantic migration: For cleaner, more reliable config handling.

Full Changelog: v1.1.0...v1.2.0

v1.1.0

Enthusiast v1.1.0

We’re excited to introduce v1.1.0, a major update packed with improvements across the board, including UI enhancements, better developer tooling, expanded plugin support, and more flexible deployment options.

Highlights

• Custom Tools Support: You can now add and configure your own tools directly through the settings.
• Improved enthusiast-common: Simplifies the process of building custom agents.
• New Plugins:
  • Google Gemini: Plug and play support for Google's latest models.
  • Ollama: Connect to Ollama-supported models in your custom deployments.
  • Shopware: Import your Shopware catalog directly to Enthusiast.
• Streamlined Deployment:
  • Starter Kit: Use Enthusiast Starter to deploy Enthusiast with custom configurations in just one click.
  • Deployment Options: Production-ready Dockerfiles
• Improved Chat Experience: Streamed agent responses, and refined submit behavior make the UI smoother and more intuitive.
• Improved Reliability: General stability improvements and bug fixes.

Full Changelog: v1.0.0...v1.1.0

New Contributors

• @dsowinski2 made their first contribution in #60

v1.0.0

docs: update internal linking