The latest articles on DEV Community by owjdie@163.com (@owjdie163com_096e40b198). https://dev.to/owjdie163com_096e40b198 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3945730%2Ff502609f-4557-4eb4-a62e-768638dda8e7.png https://dev.to/owjdie163com_096e40b198 en owjdie@163.com Fri, 22 May 2026 09:17:15 +0000 https://dev.to/owjdie163com_096e40b198/one-knowledge-base-all-ai-models-with-citations-lck https://dev.to/owjdie163com_096e40b198/one-knowledge-base-all-ai-models-with-citations-lck <p>TL;DR: I built UUMuse, an AI knowledge workspace. Upload your files once, receive answers with precise page citations. Switch LLMs freely without rebuilding context. You can also publish your knowledge base as documentation sites, embedded help widgets, APIs or MCP servers. Freshly launched, welcome to share your feedback: <a href="https://uumuse.ai" rel="noopener noreferrer">https://uumuse.ai</a></p> <p>I used to follow the same tedious routine as many others:<br> Upload PDFs and documents to ChatGPT or Claude<br> Lose all existing context while switching between different models<br> Receive plausible yet unverifiable responses<br> Repeatedly re-upload files when sharing materials with colleagues or embedding AI functions on websites<br> NotebookLM works well for basic document conversation. Still, I needed an all-in-one hub that covers memory management, AI agents, content generation and deployment, instead of a mere standalone chat tool.</p> ai productivity rag showdev