v0.2.1

Release Notes - Version 0.2.1

Improvements:

• Update translations.
• Add Claude Sonnet 3.5.
• Change default user name in chat.

Bug fix:

• Fix agent data field.
• Fix empty chunk bug in OpenAI client.

v0.2.0

Release Notes - Version 0.2.0

New Features:

• New Agent Module:
  • Added Agent module.
  • Supports voice recognition input.
  • Supports voice synthesis output.
  • Supports local large language models (including multi-modal models).
• New Workflow Nodes:
  • Multi-modal nodes: Claude-Vision, Gemini-Vision, local multi-modal, voice recognition.
  • Output nodes: audio, HTML, image rendering, table.
  • Tool nodes: screen capture, text search.
  • File processing nodes: file selection.
  • Image generation node: Dall·E.
  • Large language model nodes: Qwen, Baichuan, Deepseek, Gemini, Groq, 01ai, Minimax.

Improvements:

• UI/UX Enhancements:
  • Support for window pinning.
  • Records window size and position.
  • Added an initial setup modal for new user configuration.
  • Markdown text is now selectable.
  • Added copyable code blocks.
• Feature Updates:
  • Support for selecting the official domain name of Vector Context.
  • Support for setting hotkeys to launch Agent.
  • Updated method for handling multiple file inputs.
  • Support for debugging workflow records.
• Performance and Optimization:
  • Overall back-end structure adjustment to minimize coupling.
  • Added caching in the configuration module.
  • The vector database now processes tasks in the background.
  • Workflow task queue switched to discache deque.
• Translation Updates:
  • Continuous updates and improvements to translations.

新特性:

• 新增 Agent 模块:
  • 新增 Agent 功能。
  • 支持语音识别方式输入。
  • 支持语音合成方式输出。
  • 支持本地大语言模型（支持多模态模型）。
• 新增工作流节点:
  • 多模态节点：Claude-Vision、Gemini-Vision、本地多模态、语音识别。
  • 输出节点：音频、HTML、图片渲染、表格。
  • 工具节点：屏幕截图、文本搜索。
  • 文件处理节点：选择文件。
  • 图像生成节点：Dall·E。
  • 大语言模型节点：通义千问、百川大模型、Deepseek、Gemini、Groq、零一万物、Minimax。

改进:

• UI/UX增强:
  • 支持窗口置顶。
  • 记录窗口大小及位置。
  • 增加了初始设置模态框，以便新用户配置。
  • Markdown 正文现在可选中。
  • 增加了可复制的代码块。
• 功能更新:
  • 支持选择向量脉络官网域名。
  • 支持设定快捷键启动 Agent。
  • 更新了处理多文件输入的方法。
  • 支持对工作流记录进行调试。
• 性能和优化:
  • 后端整体结构调整，尽量解耦。
  • 在配置模块中增加了缓存。
  • 矢量数据库现在使用后台任务进行处理。
  • 工作流任务队列切换为 discache deque。
• 翻译更新:
  • 持续更新和改进翻译。

v0.1.5

Release Notes - Version 0.1.5

Bug fix:

• Fix issue where importing local model nodes fails without local models in settings.
• Fix local llm node model id bug.
• Fix translations.

v0.1.4

Release Notes - Version 0.1.4

New Features:

• Add collapse for nodes with large fields.

Bug fix:

• Fix request bug.
• Fix workflow dag detection.

v0.1.3

Release Notes - Version 0.1.3

Bug fix:

• FIx isolated nodes detection bug.

v0.1.2

Release Notes - Version 0.1.2

Bug fix:

• FIx saving bug when triggers are connected.

v0.1.1

Release Notes - Version 0.1.1

New Features:

• Add local LLMs workflow nodes.

Bug fix:

• Several bugs fixed.

v0.1.0

Release Notes - Version 0.1.0

New Features:

• Add relational database with 3 new workflow nodes.
• Add new AI models: ChatGLM, Claude, Moonshot.
• Add GLM-Vision in media processing.
• Programming function node can output error messages, console messages and generated files.
• Diagnose workflow records to view the runtime and output results of each node.

Improvements:

• Update UI design.
• Add card view in my workflows.
• Add run time in workflow records.
• Seperated settings page.
• Refactor workflow nodes structure in frontend.
• Add timer decorator for node tasks.

v0.0.31

Release Notes - Version 0.0.31

New Features:

• Add YouTube video comment crawling feature.

v0.0.30

Release Notes - Version 0.0.30

New Features:

• Add GPT4-v node.
• Add JSON format response for OpenAI node.

Improvements:

• Update to OpenAI 1.0 API.