v0.3.0

Release date: 2026.1.23

## Highlights

Introducing UltraRAG 3.0: Reject "Black Box" Development. Make Every Line of Inference Logic Visible! UltraRAG 3.0 solves the "Last Mile" problem in RAG development, developed by THUNLP, NEUIR, OpenBMB & AI9Stars. This release represents a significant milestone with **129 pull requests** merged, bringing major enhancements to functionality, UI/UX, and system stability.

* WYSIWYG Pipeline Builder: From logic to prototype in seconds. Our dual-mode builder (Canvas + Code) syncs in real-time. Click "Build" and your static logic instantly becomes an interactive UI. No more boilerplate code!

* Pixel-Level "White-Box" Visualization: Stop guessing. The "Show Thinking" panel visualizes the entire inference trajectory—loops, branches, and tool calls. Debug bad cases instantly by comparing retrieval chunks vs. model hallucinations.

* Built-in AI Developer Assistant: Stuck on config? The embedded AI Assistant knows the framework inside out. Just use natural language to generate Pipeline configurations, optimize Prompts, or explain parameters.

* DeepResearch Engine: Powered by AgentCPM-Report , it supports "Writing-as-Reasoning." The system dynamically plans, retrieves, and deepens content to generate professional, cited reports automatically.


## What's Changed


**Demo Mode & Streaming:**
* Enable demo-mode compatibility for retriever server and Milvus backend (#145) @mssssss123
* Enable streaming generation in generation server to support demo mode (#147) @mssssss123
* Full stack support (Frontend, Backend, Client) for Demo; Upgrade CLI to v3 (#149) @mssssss123
* Support displaying "thinking process" for demo (#151) @mssssss123
* Support customizable chunk params, optimize corpus server, and add thread cleanup (#167) @mssssss123


**Knowledge Base (KB):**
* Support knowledge management for demo (#160) @mssssss123
* Support uploading multi-files and folders (#161) @mssssss123
* Support selecting KB during chat (#162) @mssssss123
* Support Chinese KB names (#212) @xhd0728
* Add KB selection validation (#194) @xhd0728
* fix append index error when set an exist collection name (#266 ) @xhd0728


**Citations & Rendering:**
* Support citation by prompt (#155) @mssssss123
* Support citation for loop-like pipelines (#156) @mssssss123
* Support citation for light research pipelines (#157) @mssssss123
* Re-implement single-step and multi-step citations logic (#173) @mssssss123
* Enhance reference rendering with collapsible unused citations and improved highlight logic (#174) @mssssss123
* Integrate KaTeX for LaTeX rendering in markdown (#171) @mssssss123


**Pipeline & Agents:**
* Support SurveyCPM pipeline (#183) @mssssss123
* Support background inference suspension & optimize generation details (#185) @mssssss123
* Enhance background task management with user ID isolation (#186) @mssssss123
* Add PaperAgent demo (#189) @mssssss123
* Add AgentCPM-Report demo (#226, #230) @Kaguya-19
* Support Fascmcp v2.14.3 (#218) @xhd0728


**AI Assistant:**
* Introduce AI Assistant for configuration (#202) @mssssss123
* Enhance AI Assistant capabilities and UI (#203, #229, #231, #233) @xhd0728 @mssssss123


**Utilities:**
* Add `extra_params` for generation server (#169) @mssssss123
* Support one-click history cleanup, .docx parsing, and multiline chat input (#172) @mssssss123
* Bundle JS and CSS for offline support (#192) @xhd0728
* Use `uv` for dependency management (#222) @xhd0728


**Logic & Backend:**
* Fix server discovery issue when `server.py` is missing (#152) @mssssss123
* Re-implement search-o1 logic (#158) @mssssss123
* Fix global params bug in client (#184) @mssssss123
* Fix multi-turn chat errors (#193) @xhd0728
* Fix SurveyCPM hard mode parameters and rendering (#205, #208) @Kaguya-19 @mssssss123
* Fix event loop and session expiration issues (#216) @xhd0728
* Fix YAML parsing errors and unicode support (#213, #224) @xhd0728
* Fix image path indexing in generation (#146) @xhd0728


**UI & Display:**
* Correctly display final step in branch (#170) @mssssss123
* Prevent data misalignment in router mode (#175) @mssssss123
* Fix user role distinction and citation display in multi-chat history (#176) @mssssss123
* Fix log display issues and upgrade MCP version (#178) @mssssss123
* Fix list format checking for global variables (#191) @mssssss123
* Fix AI assistant application error when switching config view (#228) @xhd0728
* Fix text/table rendering and copy function (#237) @xhd0728
* Fix `gpu_ids` type loading and evaluation result table display (#249) @xhd0728


**General:**
* Refine chat UI, UX, and support code highlighting (#234) @xhd0728
* Split chat and config views for better workflow (#204) @xhd0728
* Set normal view to chat & enhance session history handling (#220) @xhd0728
* Enhance document preview and UI components (#238) @mssssss123


**Layout & Components:**
* Redesign knowledge base menu (#168) and management interface (#214) @mssssss123
* Optimize KB dropdown menu & citation display (#206) @xhd0728
* Update homepage layout (#196, #197) @xhd0728
* Update Builder UI components and styles (#232) @mssssss123
* Update navbar functionality and remove status indicator (#221) @mssssss123
* Enhance background task UI (#217) @xhd0728
* Unify alert dialogs (#195) @xhd0728


**Code Review & Cleanup:**
* Comprehensive code review across multiple servers (Benchmark, Corpus, Custom, Evaluation, Generation, Prompt, Reranker, Retriever, Router, Sayhello) (#239-#248, #250, #251, #255, #257) @xhd0728


**Pipeline & Server:**
* Refactor retriever server and Milvus implementation for demo (#150) @mssssss123
* Optimize markdown rendering (#153) @xhd0728
* Refactor demo citation and history saving (#154) @mssssss123
* Auto start/stop pipeline execution (#165) @mssssss123
* Optimize reference rendering logic (#166) @mssssss123
* Rename pipelines (#235) @mssssss123
* Remove and update obsolete params/tools in retriever (#200) @mssssss123
* Clear pipelines and unused parameters (#211) @mssssss123



**Documentation & Maintenance**

* Add environment requirements and usage guide (#190) @mssssss123
* Improve example cases on the welcome page (#180) @mssssss123
* Add V3 tutorial (#236) @mssssss123
* Update README and Docker usage (#227, #252, #254, #256, #261) @mssssss123 @xhd0728
* Security: Add Dependabot configuration and security notice (#258) @xhd0728
* Deps: Bump vllm from 0.13.0 to 0.14.0 (#259) @dependabot