Skip to content

v0.5.0 Release Notes

🚀 Summary

Refly Major Update: Added 2+ open source contributors, 10+ practical templates, comprehensive upgrades to knowledge base partition management, linear conversations, mind maps, and document outline features, while optimizing canvas performance to support dozens of nodes with smooth operation, creating a more powerful AI-native creation platform!

👨‍💻 New Open Source Contributors

We've welcomed 2+ new contributors in this update. Many thanks for their contributions to Refly 🎉 🎉 🎉

📑 New Templates

We've added 10+ new templates covering training materials, mind map-generated drafts, business meeting theme readings, book breakdowns, Mermaid feature studies, social media cover designs, local deployment manual writing, recipe prompts, and more.

🌟 New Features

  • 📚 Knowledge Base Partitioning (Similar to Claude Project Capability)
    • Support for creating multiple knowledge bases to manage different knowledge and canvases in separate partitions
    • Each knowledge base supports custom prompt settings, allowing different customized experiences for different knowledge bases
    • When asking AI questions, users can select which knowledge base to mount, greatly enhancing the relevance and effectiveness of AI responses
  • 💬 New Linear Conversation Support, Perfectly Adapting to Linear Sessions and Multi-threaded Canvas Conversations
    • Support for global-level linear conversations, allowing continuous follow-up questions while maintaining context, similar to ChatGPT
    • Support for linear conversation with continuous follow-up questions on any AI response card
    • Support for adding context to selected linear conversation panels
  • 🧠 New Mind Map Capabilities
    • Now supports AI-combined context generation of NotebookLM-level mind maps
    • Mind maps support adding/deleting nodes, with Markdown-level operations for each node
    • Mind maps support one-click sharing and publishing
    • Especially suitable for book breakdowns, knowledge point organization, and similar scenarios
  • 📋 New Document Outline Capability
    • Support for displaying document table of contents outlines
  • 📊 Support for New Mermaid Rendering
    • Support for 22 types of diagrams including kanban boards, pie charts, entity relationship diagrams, and more
  • 🔍 New Preview Panel Dragging and Widescreen Capabilities
    • Support for widescreen mode in preview panels
    • Support for dragging and adjusting the position of preview panels
  • 🎬 New Video Tutorial Display in Documentation
    • Showcasing 10+ video tutorials currently recorded by the community View Tutorials
  • ✏️ Node Editing Enhancements
    • All nodes can have editable titles for easy distinction and display
    • Support for alignment guides and grid snapping for canvas nodes, greatly optimizing canvas layout experience

💫 Core Problem Optimization

  • 🛠️ Fixed issues with abnormal AI response count handling when copying canvases
  • 🐳 Support for Docker local deployment with port conflict compatibility, fixing the issue of login entry not being displayed in local deployments
  • ⏱️ Optimized AI response timeout issues, fixing screen flashing, content overlay, or error problems caused by long responses (time-consuming)
  • 🔄 New detailed error display capability for model call errors, timeouts, and request rate limits
  • 📋 Fixed the issue of only partial content being copied when duplicating nodes
  • 📍 Optimized the positioning of images, resources, documents, etc. when adding them, defaulting to viewport placement
  • 🚀 Further optimized performance for large canvases, now supporting dozens of nodes with smooth operation