https://rosetears.cn/en/tags/llm/Recent content in AI Tools & LLMs on RosetearsHugo -- gohugo.ioen-USrose@rosetears.cn (Rosetears)rose@rosetears.cn (Rosetears)Tue, 26 May 2026 21:22:33 +0800https://rosetears.cn/en/posts/archives-90-90/Tue, 26 May 2026 15:59:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-90-90/To reduce an AIGC suspicion rate, the core task is not to turn a paper into casual speech, but to first locate risky sections, then preserve academic facts while strengthening concrete context. This article provides three workflows: segmented Agent rewriting, web-based AI handling, and solidifying rules as a Skill. It is useful for graduation-thesis revision and emphasizes that terminology, data, citations, and section numbers must be manually reviewed, especially for drafts that already have a detection report.https://rosetears.cn/en/posts/archives-87-87/Thu, 26 Feb 2026 10:21:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-87-87/If you are unsure which AI coding subscription to buy, this hands-on cost comparison compares GLM, Minimax, Kimi, Codex, Gemini, and Claude. It explains quota rules, hidden limits, Kimi’s 3× expansion, GLM weekly caps, and recommended combinations for different budgets and usage levels so developers can avoid wasting money.https://rosetears.cn/en/posts/archives-86-86/Wed, 11 Feb 2026 22:15:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-86-86/This SCI paper-polishing prompt uses a roundtable-meeting format and forces the AI to simulate an editorial workflow similar to Nature. It does not rewrite immediately. Instead, a logic architect and reviewers first run three rounds of debate, then a writing specialist outputs two versions, Draft A and Draft B. The process raises information density and reduces the mechanical feel of AI-generated prose, making it suitable for detailed academic-paper revision.https://rosetears.cn/en/posts/archives-85-85/Sat, 31 Jan 2026 16:40:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-85-85/A framework for keeping long-running AI development from drifting, using tasks.md for process, feature_list.json for state, progress.txt for handoff, and reproducible validation bundles with scripts, logs, and screenshots. Worker implements and delivers a bundle; Supervisor runs acceptance and records evidence; after PASS, state and commits are archived.https://rosetears.cn/en/posts/archives-82-82/Wed, 21 Jan 2026 10:34:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-82-82/A practical explanation of Spec-Driven Development (SDD): write What/Why and acceptance criteria as a living specification before asking AI to implement How. The article explains why vibe coding easily loses intent, and shows how SpecKit (Constitution → Specify → Plan → Tasks) fits 0→1 development while OpenSpec (proposal → apply → archive) fits 1→n iteration.https://rosetears.cn/en/posts/archives-81-81/Thu, 15 Jan 2026 16:21:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-81-81/This article clarifies six often-confused AI tool concepts: Prompt handles temporary conversation instructions, Skills preserve standard workflows, MCP exposes external tools, Commands create shortcut templates, Agent runs autonomous loops, and Subagent isolates parallel tasks. It compares definitions, use cases, engineering cost, reuse, and decision criteria based on Anthropic documentation and LangChain-style agent concepts.https://rosetears.cn/en/posts/archives-77-77/Thu, 04 Dec 2025 12:56:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-77-77/This article combines AI self-improvement with practical AI prompts and collects six deep self-questioning frameworks suitable for AI products with global memory: a brutal life-truth scanner, HUMAN 3.0 evaluation, ruthless debate partner, letter-writing decompression, legends versus ordinary people questions, and using AI to kill procrastination. It helps you use AI to see yourself more clearly and upgrade your lifestyle.https://rosetears.cn/en/posts/archives-63-63/Wed, 01 Oct 2025 16:50:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-63-63/A step-by-step guide to bypassing the Claude Code extension login check and routing the plugin through Claude Code Router so it can connect to large-model services of your choice.https://rosetears.cn/en/posts/archives-62-62/Wed, 10 Sep 2025 19:53:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-62-62/A Zotero PDF translation guide based on zotero-pdf2zh and PDFMathTranslate Next, covering Python and Docker deployments, plugin configuration, formula and figure preservation, and bilingual reading.https://rosetears.cn/en/posts/archives-61-61/Wed, 27 Aug 2025 20:24:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-61-61/A Claude Code Router tutorial for Gemini, Kimi, DeepSeek, GLM, and Qwen, covering CCR configuration, PROXY_URL, Anthropic versus OpenAI endpoints, /model switching, and common errors.https://rosetears.cn/en/posts/archives-60-60/Fri, 22 Aug 2025 21:27:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-60-60/A hands-on VS Code and Claude Code Router guide: configure CCR, set API keys in ccr ui, set ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN, then launch Claude inside VS Code.https://rosetears.cn/en/posts/archives-56-56/Wed, 23 Jul 2025 10:20:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-56-56/A practical prompt-engineering guide that introduces a powerful meta-prompt for turning AI into a strategic advisor that clarifies requirements, asks better questions, and produces more reliable outputs.https://rosetears.cn/en/posts/archives-53-53/Thu, 05 Jun 2025 10:57:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-53-53/A beginner-friendly guide to configuring PDF2zh 2.0 and the Zotero plugin through Windows .exe, local uv, and Docker deployment, helping you build an efficient research-translation workflow.https://rosetears.cn/en/posts/archives-52-52/Mon, 26 May 2025 16:21:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-52-52/Use multimodal models such as GPT-4o with three reusable prompts to load PDFs, summarize article structure, generate deep-reading notes, and produce mind-map-ready references for research reading.https://rosetears.cn/en/posts/archives-46-46/Mon, 19 May 2025 19:19:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-46-46/Explore Bohrium AI Search for second-level retrieval across 160 million papers, plus knowledge bases, notebooks, subscriptions, and other research-cloud functions that improve discovery and research efficiency.https://rosetears.cn/en/posts/archives-45-45/Wed, 23 Apr 2025 20:01:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-45-45/Metaso AI Search, built by Shanghai Metaso Network Technology, combines its MetaLLM model with an ad-free structured AI teacher. It can generate step-by-step explanations, slides, mind maps, and multimodal outputs for study, research, and work.https://rosetears.cn/en/posts/archives-44-44/Tue, 22 Apr 2025 10:41:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-44-44/This article explains how to download, install, and configure the Cherry Studio AI assistant desktop client on Windows, covering LLM service setup and common shortcuts so you can get started quickly and use it efficiently.https://rosetears.cn/en/posts/archives-28-28/Thu, 13 Feb 2025 21:07:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-28-28/This article introduces an efficient polishing prompt based on large language models. It improves the writing experience, provides a quick-use option, and offers model-selection suggestions for more efficient content creation.https://rosetears.cn/en/posts/archives-18-18/Tue, 11 Feb 2025 20:34:00 +0800rose@rosetears.cn (Rosetears)https://rosetears.cn/en/posts/archives-18-18/This blog post introduces how to use AI tools such as DeepSeek, Tongyi Qianwen, and Kimi to quickly generate PPTs and edit them. By combining AI text generation with automated PPT creation, it improves work efficiency and is suitable for students and working professionals.