[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-llama-farm--llamafarm":3,"tool-llama-farm--llamafarm":64},[4,17,27,35,44,52],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",140436,2,"2026-04-05T23:32:43",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":10,"last_commit_at":41,"category_tags":42,"status":16},4292,"Deep-Live-Cam","hacksider\u002FDeep-Live-Cam","Deep-Live-Cam 是一款专注于实时换脸与视频生成的开源工具，用户仅需一张静态照片，即可通过“一键操作”实现摄像头画面的即时变脸或制作深度伪造视频。它有效解决了传统换脸技术流程繁琐、对硬件配置要求极高以及难以实时预览的痛点，让高质量的数字内容创作变得触手可及。\n\n这款工具不仅适合开发者和技术研究人员探索算法边界，更因其极简的操作逻辑（仅需三步：选脸、选摄像头、启动），广泛适用于普通用户、内容创作者、设计师及直播主播。无论是为了动画角色定制、服装展示模特替换，还是制作趣味短视频和直播互动，Deep-Live-Cam 都能提供流畅的支持。\n\n其核心技术亮点在于强大的实时处理能力，支持口型遮罩（Mouth Mask）以保留使用者原始的嘴部动作，确保表情自然精准；同时具备“人脸映射”功能，可同时对画面中的多个主体应用不同面孔。此外，项目内置了严格的内容安全过滤机制，自动拦截涉及裸露、暴力等不当素材，并倡导用户在获得授权及明确标注的前提下合规使用，体现了技术发展与伦理责任的平衡。",88924,"2026-04-06T03:28:53",[13,14,15,43],"视频",{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":23,"last_commit_at":50,"category_tags":51,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":53,"name":54,"github_repo":55,"description_zh":56,"stars":57,"difficulty_score":23,"last_commit_at":58,"category_tags":59,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,60,43,61,15,62,26,13,63],"数据工具","插件","其他","音频",{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":80,"owner_twitter":79,"owner_website":81,"owner_url":82,"languages":83,"stars":121,"forks":122,"last_commit_at":123,"license":124,"difficulty_score":23,"env_os":125,"env_gpu":126,"env_ram":127,"env_deps":128,"category_tags":140,"github_topics":141,"view_count":23,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":161,"updated_at":162,"faqs":163,"releases":194},4131,"llama-farm\u002Fllamafarm","llamafarm","Deploy any AI model, agent, database, RAG, and pipeline locally or remotely in minutes","LlamaFarm 是一个开源的本地化 AI 平台，旨在让用户无需依赖云端服务，即可在个人硬件上快速部署大模型、智能体、数据库及 RAG（检索增强生成）应用。它主要解决了企业对数据隐私的担忧以及使用云端 API 产生的高昂费用问题，确保所有数据处理均在本地完成，即使断网也能正常运行。\n\n无论是希望保护敏感数据的企业管理者、需要离线开发环境的开发者，还是想低成本体验前沿 AI 技术的普通用户，都能从中受益。LlamaFarm 提供了直观的桌面应用程序，无需编写代码即可上手；同时也支持命令行和源码模式，满足深度定制需求。\n\n其技术亮点在于强大的硬件自适应能力，能自动利用 Apple Silicon、NVIDIA 或 AMD 的 GPU\u002FNPU 进行加速。功能方面，它不仅支持文档问答、自定义分类训练和异常检测，还集成了 OCR 文字提取、命名实体识别以及通过 MCP 协议连接外部工具的能力。用户可以在 Ollama、vLLM 等多种运行时之间自由切换，轻松构建私有化的 AI 工作流。","# LlamaFarm - Edge AI for Everyone\n\n> Enterprise AI capabilities on your own hardware. No cloud required.\n\n[![License: Apache 2.0](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fllama-farm\u002Fllamafarm)](LICENSE)\n[![Python 3.10+](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.10+-blue.svg)](https:\u002F\u002Fwww.python.org\u002Fdownloads\u002F)\n[![Go 1.24+](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgo-1.24+-00ADD8.svg)](https:\u002F\u002Fgo.dev\u002Fdl\u002F)\n[![Docs](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdocs-latest-4C51BF.svg)](https:\u002F\u002Fdocs.llamafarm.dev)\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F1392890421771899026.svg)](https:\u002F\u002Fdiscord.gg\u002FRrAUXTCVNF)\n\n**LlamaFarm** is an open-source AI platform that runs entirely on your hardware. Build RAG applications, train custom classifiers, detect anomalies, and run document processing—all locally with complete privacy.\n\n- 🔒 **Complete Privacy** — Your data never leaves your device\n- 💰 **No API Costs** — Use open-source models without per-token fees\n- 🌐 **Offline Capable** — Works without internet once models are downloaded\n- ⚡ **Hardware Optimized** — Automatic GPU\u002FNPU acceleration on Apple Silicon, NVIDIA, and AMD\n\n### Desktop App Downloads\n\nGet started instantly — no command line required:\n\n| Platform | Download |\n|----------|----------|\n| **Mac (Universal)** | [Download](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-mac-universal.dmg) |\n| **Windows** | [Download](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-windows.exe) |\n| **Linux (x86_64)** | [Download](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-linux-x86_64.AppImage) |\n| **Linux (ARM64)** | [Download](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-linux-arm64.AppImage) |\n\n---\n\n### What Can You Build?\n\n| Capability | Description |\n|-----------|-------------|\n| **RAG (Retrieval-Augmented Generation)** | Ingest PDFs, docs, CSVs and query them with AI |\n| **Custom Classifiers** | Train text classifiers with 8-16 examples using SetFit |\n| **Anomaly Detection** | 12+ algorithms for batch and streaming anomaly detection |\n| **Tool Calling (MCP)** | Connect models to external tools via Model Context Protocol |\n| **OCR & Document Extraction** | Extract text and structured data from images and PDFs |\n| **Named Entity Recognition** | Find people, organizations, and locations |\n| **Multi-Model Runtime** | Switch between Ollama, OpenAI, vLLM, or local GGUF models |\n\n**Video demo (90 seconds):** https:\u002F\u002Fyoutu.be\u002FW7MHGyN0MdQ\n\n---\n\n## Quickstart\n\n### Option 1: Desktop App\n\nDownload the desktop app above and run it. No additional setup required.\n\n### Option 2: CLI + Development Mode\n\n1. **Install the CLI**\n\n   macOS \u002F Linux:\n   ```bash\n   curl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fllama-farm\u002Fllamafarm\u002Fmain\u002Finstall.sh | bash\n   ```\n\n   Windows (PowerShell):\n   ```powershell\n   irm https:\u002F\u002Fraw.githubusercontent.com\u002Fllama-farm\u002Fllamafarm\u002Fmain\u002Finstall.ps1 | iex\n   ```\n\n   Or download directly from [releases](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest).\n\n2. **Create and run a project**\n\n   ```bash\n   lf init my-project      # Generates llamafarm.yaml\n   lf start                # Starts services and opens Designer UI\n   ```\n\n3. **Chat with your AI**\n\n   ```bash\n   lf chat                           # Interactive chat\n   lf chat \"Hello, LlamaFarm!\"       # One-off message\n   ```\n\nThe Designer web interface is available at `http:\u002F\u002Flocalhost:14345`.\n\n### Option 3: Development from Source\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm.git\ncd llamafarm\n\n# Install Nx globally and initialize the workspace\nnpm install -g nx\nnx init --useDotNxInstallation --interactive=false  # Required on first clone\n\n# Start all services (run each in a separate terminal)\nnx start server           # FastAPI server (port 14345)\nnx start rag              # RAG worker for document processing\nnx start universal-runtime # ML models, OCR, embeddings (port 11540)\n```\n\n---\n\n## Architecture\n\nLlamaFarm consists of three main services:\n\n| Service | Port | Purpose |\n|---------|------|---------|\n| **Server** | 14345 | FastAPI REST API, Designer web UI, project management |\n| **RAG Worker** | - | Celery worker for async document processing |\n| **Universal Runtime** | 11540 | ML model inference, embeddings, OCR, anomaly detection |\n\nAll configuration lives in `llamafarm.yaml`—no scattered settings or hidden defaults.\n\n---\n\n## Runtime Options\n\n### Universal Runtime (Recommended)\n\nThe Universal Runtime provides access to HuggingFace models plus specialized ML capabilities:\n\n- **Text Generation** - Any HuggingFace text model\n- **Embeddings** - sentence-transformers and other embedding models\n- **OCR** - Text extraction from images\u002FPDFs (Surya, EasyOCR, PaddleOCR, Tesseract)\n- **Document Extraction** - Forms, invoices, receipts via vision models\n- **Text Classification** - Pre-trained or custom models via SetFit\n- **Named Entity Recognition** - Extract people, organizations, locations\n- **Reranking** - Cross-encoder models for improved RAG quality\n- **Anomaly Detection** - Isolation Forest, One-Class SVM, Local Outlier Factor, Autoencoders\n\n```yaml\nruntime:\n  models:\n    default:\n      provider: universal\n      model: Qwen\u002FQwen2.5-1.5B-Instruct\n      base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n```\n\n### Ollama\n\nSimple setup for GGUF models with CPU\u002FGPU acceleration:\n\n```yaml\nruntime:\n  models:\n    default:\n      provider: ollama\n      model: qwen3:8b\n      base_url: http:\u002F\u002Flocalhost:11434\u002Fv1\n```\n\n### OpenAI-Compatible\n\nWorks with vLLM, Together, Mistral API, or any OpenAI-compatible endpoint:\n\n```yaml\nruntime:\n  models:\n    default:\n      provider: openai\n      model: gpt-4o\n      base_url: https:\u002F\u002Fapi.openai.com\u002Fv1\n      api_key: ${OPENAI_API_KEY}\n```\n\n---\n\n## Core Workflows\n\n### CLI Commands\n\n| Task | Command |\n|------|---------|\n| Initialize project | `lf init my-project` |\n| Start services | `lf start` |\n| Interactive chat | `lf chat` |\n| One-off message | `lf chat \"Your question\"` |\n| List models | `lf models list` |\n| Use specific model | `lf chat --model powerful \"Question\"` |\n| Create dataset | `lf datasets create -s pdf_ingest -b main_db research` |\n| Upload files (auto-process by default) | `lf datasets upload research .\u002Fdocs\u002F*.pdf` |\n| Process dataset (if you skipped auto-process) | `lf datasets process research` |\n| Query RAG | `lf rag query --database main_db \"Your query\"` |\n| Check RAG health | `lf rag health` |\n\n### RAG Pipeline\n\n1. **Create a dataset** linked to a processing strategy and database\n2. **Upload files** (PDF, DOCX, Markdown, TXT) — processing runs automatically unless you pass `--no-process`\n3. **Process manually** only when you intentionally skipped auto-processing (e.g., large batches)\n4. **Query** using semantic search with optional metadata filtering\n\n```bash\nlf datasets create -s default -b main_db research\nlf datasets upload research .\u002Fpapers\u002F*.pdf                 # auto-processes by default\n# For large batches:\n# lf datasets upload research .\u002Fpapers\u002F*.pdf --no-process\n# lf datasets process research\nlf rag query --database main_db \"What are the key findings?\"\n```\n\n### Designer Web UI\n\nThe Designer at `http:\u002F\u002Flocalhost:14345` provides:\n\n- **Project management** with briefs and quick actions\n- **Visual dataset management** with drag-and-drop uploads\n- **Database & RAG configuration** with built-in query testing\n- **Prompt engineering** with template variables and testing\n- **Interactive chat** with RAG toggle and retrieved context display\n- **Config editor** with syntax highlighting, validation, and auto-completion\n- Switch between visual Designer and raw YAML modes in any section\n\nSee the [Designer Features Guide](docs\u002Fwebsite\u002Fdocs\u002Fdesigner\u002Ffeatures.md) for details.\n\n---\n\n## Configuration\n\n`llamafarm.yaml` is the source of truth for each project:\n\n```yaml\nversion: v1\nname: my-assistant\nnamespace: default\n\n# Multi-model configuration\nruntime:\n  default_model: fast\n\n  models:\n    fast:\n      description: \"Fast local model\"\n      provider: universal\n      model: Qwen\u002FQwen2.5-1.5B-Instruct\n      base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n\n    powerful:\n      description: \"More capable model\"\n      provider: universal\n      model: Qwen\u002FQwen2.5-7B-Instruct\n      base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n\n# System prompts\nprompts:\n  - name: default\n    messages:\n      - role: system\n        content: You are a helpful assistant.\n\n# RAG configuration\nrag:\n  databases:\n    - name: main_db\n      type: ChromaStore\n      default_embedding_strategy: default_embeddings\n      default_retrieval_strategy: semantic_search\n      embedding_strategies:\n        - name: default_embeddings\n          type: UniversalEmbedder\n          config:\n            model: sentence-transformers\u002Fall-MiniLM-L6-v2\n            base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n      retrieval_strategies:\n        - name: semantic_search\n          type: BasicSimilarityStrategy\n          config:\n            top_k: 5\n\n  data_processing_strategies:\n    - name: default\n      parsers:\n        - type: PDFParser_LlamaIndex\n          config:\n            chunk_size: 1000\n            chunk_overlap: 100\n        - type: MarkdownParser_Python\n          config:\n            chunk_size: 1000\n      extractors: []\n\n# Dataset definitions\ndatasets:\n  - name: research\n    data_processing_strategy: default\n    database: main_db\n```\n\n### Environment Variable Substitution\n\nUse `${VAR}` syntax to inject secrets from `.env` files:\n\n```yaml\nruntime:\n  models:\n    openai:\n      api_key: ${OPENAI_API_KEY}\n      # With default: ${OPENAI_API_KEY:-sk-default}\n      # From specific file: ${file:.env.production:API_KEY}\n```\n\nSee the [Configuration Guide](docs\u002Fwebsite\u002Fdocs\u002Fconfiguration\u002Findex.md) for complete reference.\n\n---\n\n## REST API\n\nLlamaFarm provides an OpenAI-compatible REST API:\n\n**Chat Completions**\n```bash\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fprojects\u002Fdefault\u002Fmy-project\u002Fchat\u002Fcompletions \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\n    \"messages\": [{\"role\": \"user\", \"content\": \"Hello\"}],\n    \"stream\": false,\n    \"rag_enabled\": true\n  }'\n```\n\n**RAG Query**\n```bash\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fprojects\u002Fdefault\u002Fmy-project\u002Frag\u002Fquery \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\n    \"query\": \"What are the requirements?\",\n    \"database\": \"main_db\",\n    \"top_k\": 5\n  }'\n```\n\nSee the [API Reference](docs\u002Fwebsite\u002Fdocs\u002Fapi\u002Findex.md) for all endpoints.\n\n---\n\n## Specialized ML Capabilities\n\nThe Universal Runtime provides endpoints beyond chat:\n\n### OCR & Document Extraction\n\n```bash\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fvision\u002Focr \\\n  -F \"file=@document.pdf\" \\\n  -F \"model=surya\"\n```\n\n### Anomaly Detection\n\nLlamaFarm supports 12+ anomaly detection algorithms via PyOD, with both batch and streaming modes.\n\n```bash\n# Train on normal data\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fml\u002Fanomaly\u002Ffit \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"model\": \"sensor-detector\", \"backend\": \"ecod\", \"data\": [[22.1], [23.5], ...]}'\n\n# Detect anomalies\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fml\u002Fanomaly\u002Fdetect \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"model\": \"sensor-detector\", \"data\": [[22.0], [100.0], [23.0]], \"threshold\": 0.5}'\n\n# Streaming detection (handles cold start, auto-retraining, sliding windows)\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fml\u002Fanomaly\u002Fstream \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"model\": \"live-sensor\", \"data\": {\"temperature\": 72.5}, \"backend\": \"ecod\"}'\n```\n\n**Available backends:** `ecod` (recommended), `isolation_forest`, `one_class_svm`, `local_outlier_factor`, `autoencoder`, `hbos`, `copod`, `knn`, `mcd`, `cblof`, `suod`, `loda`\n\n### Text Classification & NER\n\nSee the [Models Guide](docs\u002Fwebsite\u002Fdocs\u002Fmodels\u002Findex.md) for complete documentation.\n\n### Tool Calling (MCP)\n\nGive models access to external tools via the Model Context Protocol:\n\n```yaml\n# In llamafarm.yaml\nmcp:\n  servers:\n    - name: filesystem\n      transport: stdio\n      command: npx\n      args: ['-y', '@modelcontextprotocol\u002Fserver-filesystem', '\u002Fdata']\n\nruntime:\n  models:\n    - name: assistant\n      provider: ollama\n      model: llama3.1:8b\n      mcp_servers: [filesystem]\n```\n\nLlamaFarm also exposes its own API as MCP tools for use with Claude Desktop, Cursor, and other MCP clients. See the [Tool Calling Guide](docs\u002Fwebsite\u002Fdocs\u002Fmcp\u002Findex.md).\n\n---\n\n## Examples\n\n| Example | Description | Location |\n|---------|-------------|----------|\n| **RAG Examples** | | |\n| Large Complex PDFs | Multi-megabyte planning ordinances | `examples\u002Flarge_complex_rag\u002F` |\n| Many Small Files | FDA correspondence letters | `examples\u002Fmany_small_file_rag\u002F` |\n| Mixed Formats | PDF, Markdown, HTML, text, and code | `examples\u002Fmixed_format_rag\u002F` |\n| Quick Notes | Rapid smoke tests with small files | `examples\u002Fquick_rag\u002F` |\n| **Anomaly Detection** | | |\n| Quick Start | Simplest anomaly detection example | `examples\u002Fanomaly\u002F01_quick_start.py` |\n| Fraud Detection | Training, saving, loading models | `examples\u002Fanomaly\u002F02_fraud_detection.py` |\n| Streaming Sensors | IoT monitoring with rolling features | `examples\u002Fanomaly\u002F03_streaming_sensors.py` |\n| Backend Comparison | Compare all 12 algorithms | `examples\u002Fanomaly\u002F04_backend_comparison.py` |\n| **Use Cases** | | |\n| FDA Letters Assistant | Regulatory document analysis | `examples\u002Ffda_rag\u002F` |\n| Government Planning | Large ordinance documents | `examples\u002Fgov_rag\u002F` |\n\nSee [`examples\u002FREADME.md`](examples\u002FREADME.md) for setup instructions and the full list.\n\n---\n\n## Industry Use Cases\n\nLlamaFarm is used across industries for document analysis, monitoring, and fraud detection:\n\n- **[Pharmaceutical & Therapeutics](docs\u002Fwebsite\u002Fdocs\u002Fuse-cases\u002Fpharmaceutical-fda.md)** — Analyze FDA correspondence, track regulatory questions\n- **[IoT Sensor Monitoring](docs\u002Fwebsite\u002Fdocs\u002Fuse-cases\u002Fiot-sensor-monitoring.md)** — Real-time streaming anomaly detection with automatic retraining\n- **[Financial Fraud Detection](docs\u002Fwebsite\u002Fdocs\u002Fuse-cases\u002Ffinancial-fraud-detection.md)** — Multi-stage fraud detection with velocity and behavioral patterns\n\n---\n\n## Development & Testing\n\n```bash\n# Python server tests\ncd server && uv sync && uv run --group test python -m pytest\n\n# CLI tests\ncd cli && go test .\u002F...\n\n# RAG tests\ncd rag && uv sync && uv run pytest tests\u002F\n\n# Universal Runtime tests\ncd runtimes\u002Funiversal && uv sync && uv run pytest tests\u002F\n\n# Build docs\nnx build docs\n```\n\n---\n\n## Extensibility\n\n- **Add runtimes** by implementing provider support and updating schema\n- **Add vector stores** by implementing store backends (Chroma, Qdrant, etc.)\n- **Add parsers** for new file formats (PDF, DOCX, HTML, CSV, etc.)\n- **Add extractors** for custom metadata extraction\n- **Add CLI commands** under `cli\u002Fcmd\u002F`\n\nSee the [Extending Guide](docs\u002Fwebsite\u002Fdocs\u002Fextending\u002Findex.md) for step-by-step instructions.\n\n---\n\n## Community & Support\n\n- [Discord](https:\u002F\u002Fdiscord.gg\u002FRrAUXTCVNF) - Chat with the team and community\n- [GitHub Issues](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues) - Bug reports and feature requests\n- [Discussions](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fdiscussions) - Ideas and proposals\n- [Contributing Guide](CONTRIBUTING.md) - Code style and contribution process\n\n---\n\n## License\n\nLicensed under the [Apache 2.0 License](LICENSE). See [CREDITS](CREDITS.md) for acknowledgments.\n\n---\n\nBuild locally. Deploy anywhere. Own your AI.\n","# LlamaFarm - 面向每个人的边缘人工智能\n\n> 在您自己的硬件上实现企业级 AI 功能。无需云端。\n\n[![许可证：Apache 2.0](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fllama-farm\u002Fllamafarm)](LICENSE)\n[![Python 3.10+](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.10+-blue.svg)](https:\u002F\u002Fwww.python.org\u002Fdownloads\u002F)\n[![Go 1.24+](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fgo-1.24+-00ADD8.svg)](https:\u002F\u002Fgo.dev\u002Fdl\u002F)\n[![文档](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdocs-latest-4C51BF.svg)](https:\u002F\u002Fdocs.llamafarm.dev)\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F1392890421771899026.svg)](https:\u002F\u002Fdiscord.gg\u002FRrAUXTCVNF)\n\n**LlamaFarm** 是一个完全在您的硬件上运行的开源 AI 平台。您可以构建 RAG 应用程序、训练自定义分类器、检测异常并进行文档处理——所有这些都在本地完成，确保完全的隐私。\n\n- 🔒 **完全隐私** — 您的数据永远不会离开您的设备\n- 💰 **无 API 费用** — 使用开源模型，无需按 token 计费\n- 🌐 **离线可用** — 下载模型后即可在没有互联网的情况下工作\n- ⚡ **硬件优化** — 自动支持 Apple Silicon、NVIDIA 和 AMD 的 GPU\u002FNPU 加速\n\n### 桌面应用下载\n\n立即开始使用，无需命令行：\n\n| 平台 | 下载 |\n|----------|----------|\n| **Mac (通用)** | [下载](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-mac-universal.dmg) |\n| **Windows** | [下载](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-windows.exe) |\n| **Linux (x86_64)** | [下载](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-linux-x86_64.AppImage) |\n| **Linux (ARM64)** | [下载](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-linux-arm64.AppImage) |\n\n---\n\n### 您可以构建什么？\n\n| 功能 | 描述 |\n|-----------|-------------|\n| **RAG（检索增强生成）** | 导入 PDF、文档、CSV 文件，并通过 AI 进行查询 |\n| **自定义分类器** | 使用 SetFit，仅需 8–16 个示例即可训练文本分类器 |\n| **异常检测** | 12 种以上的算法，适用于批量和流式异常检测 |\n| **工具调用（MCP）** | 通过模型上下文协议将模型连接到外部工具 |\n| **OCR & 文档提取** | 从图像和 PDF 中提取文本和结构化数据 |\n| **命名实体识别** | 查找人物、组织和地点 |\n| **多模型运行时** | 可在 Ollama、OpenAI、vLLM 或本地 GGUF 模型之间切换 |\n\n**视频演示（90 秒）：** https:\u002F\u002Fyoutu.be\u002FW7MHGyN0MdQ\n\n---\n\n## 快速入门\n\n### 选项 1：桌面应用\n\n下载上述桌面应用并运行即可。无需额外设置。\n\n### 选项 2：CLI + 开发模式\n\n1. **安装 CLI**\n\n   macOS \u002F Linux：\n   ```bash\n   curl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fllama-farm\u002Fllamafarm\u002Fmain\u002Finstall.sh | bash\n   ```\n\n   Windows（PowerShell）：\n   ```powershell\n   irm https:\u002F\u002Fraw.githubusercontent.com\u002Fllama-farm\u002Fllamafarm\u002Fmain\u002Finstall.ps1 | iex\n   ```\n\n   或直接从 [发布页面](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest) 下载。\n\n2. **创建并运行项目**\n\n   ```bash\n   lf init my-project      # 生成 llamafarm.yaml\n   lf start                # 启动服务并打开 Designer UI\n   ```\n\n3. **与您的 AI 对话**\n\n   ```bash\n   lf chat                           # 交互式聊天\n   lf chat \"Hello, LlamaFarm!\"       # 单次消息\n   ```\n\nDesigner 网页界面可在 `http:\u002F\u002Flocalhost:14345` 访问。\n\n### 选项 3：从源代码开发\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm.git\ncd llamafarm\n\n# 全局安装 Nx 并初始化工作区\nnpm install -g nx\nnx init --useDotNxInstallation --interactive=false  # 第一次克隆时必须执行\n\n# 启动所有服务（每个服务在一个单独的终端中运行）\nnx start server           # FastAPI 服务器（端口 14345）\nnx start rag              # 用于文档处理的 RAG 工作进程\nnx start universal-runtime # 机器学习模型、OCR、嵌入等（端口 11540）\n```\n\n---\n\n## 架构\n\nLlamaFarm 由三个主要服务组成：\n\n| 服务 | 端口 | 用途 |\n|---------|------|---------|\n| **服务器** | 14345 | FastAPI REST API、Designer 网页界面、项目管理 |\n| **RAG 工作进程** | - | Celery 工作进程，用于异步文档处理 |\n| **通用运行时** | 11540 | 机器学习模型推理、嵌入、OCR、异常检测 |\n\n所有配置都存储在 `llamafarm.yaml` 中，不存在分散的设置或隐藏的默认值。\n\n---\n\n## 运行时选项\n\n### 通用运行时（推荐）\n\n通用运行时提供对 HuggingFace 模型以及专用机器学习功能的访问：\n\n- **文本生成** - 任何 HuggingFace 文本模型\n- **嵌入** - sentence-transformers 等嵌入模型\n- **OCR** - 从图像\u002FPDF 中提取文本（Surya、EasyOCR、PaddleOCR、Tesseract）\n- **文档提取** - 通过视觉模型提取表格、发票、收据等\n- **文本分类** - 通过 SetFit 使用预训练或自定义模型\n- **命名实体识别** - 提取人物、组织、地点\n- **重排序** - 使用交叉编码器提高 RAG 质量\n- **异常检测** - 隔离森林、单类 SVM、局部异常因子、自编码器\n\n```yaml\nruntime:\n  models:\n    default:\n      provider: universal\n      model: Qwen\u002FQwen2.5-1.5B-Instruct\n      base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n```\n\n### Ollama\n\n简单设置，支持 GGUF 模型，并具备 CPU\u002FGPU 加速：\n\n```yaml\nruntime:\n  models:\n    default:\n      provider: ollama\n      model: qwen3:8b\n      base_url: http:\u002F\u002Flocalhost:11434\u002Fv1\n```\n\n### OpenAI 兼容\n\n可与 vLLM、Together、Mistral API 或任何 OpenAI 兼容的端点配合使用：\n\n```yaml\nruntime:\n  models:\n    default:\n      provider: openai\n      model: gpt-4o\n      base_url: https:\u002F\u002Fapi.openai.com\u002Fv1\n      api_key: ${OPENAI_API_KEY}\n```\n\n---\n\n## 核心工作流程\n\n### CLI 命令\n\n| 任务 | 命令 |\n|------|---------|\n| 初始化项目 | `lf init my-project` |\n| 启动服务 | `lf start` |\n| 交互式聊天 | `lf chat` |\n| 单次消息 | `lf chat \"Your question\"` |\n| 列出模型 | `lf models list` |\n| 使用特定模型 | `lf chat --model powerful \"Question\"` |\n| 创建数据集 | `lf datasets create -s pdf_ingest -b main_db research` |\n| 上传文件（默认自动处理） | `lf datasets upload research .\u002Fdocs\u002F*.pdf` |\n| 手动处理数据集（如果跳过了自动处理） | `lf datasets process research` |\n| 查询 RAG | `lf rag query --database main_db \"Your query\"` |\n| 检查 RAG 健康状况 | `lf rag health` |\n\n### RAG 流程\n\n1. **创建数据集**，关联处理策略和数据库\n2. **上传文件**（PDF、DOCX、Markdown、TXT）——除非您指定 `--no-process`，否则会自动处理\n3. **手动处理** 仅在您有意跳过自动处理时进行（例如大批量文件）\n4. **查询** 使用语义搜索，并可选择性地过滤元数据\n\n```bash\nlf datasets create -s default -b main_db research\nlf datasets upload research .\u002Fpapers\u002F*.pdf                 # 默认自动处理\n\n# 对于大型批次：\n# lf 数据集上传 research .\u002Fpapers\u002F*.pdf --no-process\n# lf 数据集处理 research\nlf rag 查询 --database main_db \"主要发现是什么？\"\n```\n\n### Designer Web UI\n\nDesigner 的地址为 `http:\u002F\u002Flocalhost:14345`，提供以下功能：\n\n- **项目管理**，包括简报和快速操作\n- **可视化数据集管理**，支持拖放上传\n- **数据库与 RAG 配置**，内置查询测试功能\n- **提示工程**，支持模板变量和测试\n- **交互式聊天**，可切换 RAG 模式并显示检索到的上下文\n- **配置编辑器**，具备语法高亮、验证和自动补全功能\n- 在任何部分均可在可视化 Designer 和原始 YAML 模式之间切换\n\n详细信息请参阅 [Designer 功能指南](docs\u002Fwebsite\u002Fdocs\u002Fdesigner\u002Ffeatures.md)。\n\n---\n\n## 配置\n\n`llamafarm.yaml` 是每个项目的权威配置文件：\n\n```yaml\nversion: v1\nname: my-assistant\nnamespace: default\n\n# 多模型配置\nruntime:\n  default_model: fast\n\n  models:\n    fast:\n      description: \"快速本地模型\"\n      provider: universal\n      model: Qwen\u002FQwen2.5-1.5B-Instruct\n      base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n\n    powerful:\n      description: \"更强大的模型\"\n      provider: universal\n      model: Qwen\u002FQwen2.5-7B-Instruct\n      base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n\n# 系统提示\nprompts:\n  - name: default\n    messages:\n      - role: system\n        content: 你是一位乐于助人的助手。\n\n# RAG 配置\nrag:\n  databases:\n    - name: main_db\n      type: ChromaStore\n      default_embedding_strategy: default_embeddings\n      default_retrieval_strategy: semantic_search\n      embedding_strategies:\n        - name: default_embeddings\n          type: UniversalEmbedder\n          config:\n            model: sentence-transformers\u002Fall-MiniLM-L6-v2\n            base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n      retrieval_strategies:\n        - name: semantic_search\n          type: BasicSimilarityStrategy\n          config:\n            top_k: 5\n\n  data_processing_strategies:\n    - name: default\n      parsers:\n        - type: PDFParser_LlamaIndex\n          config:\n            chunk_size: 1000\n            chunk_overlap: 100\n        - type: MarkdownParser_Python\n          config:\n            chunk_size: 1000\n      extractors: []\n\n# 数据集定义\ndatasets:\n  - name: research\n    data_processing_strategy: default\n    database: main_db\n```\n\n### 环境变量替换\n\n使用 `${VAR}` 语法从 `.env` 文件中注入密钥：\n\n```yaml\nruntime:\n  models:\n    openai:\n      api_key: ${OPENAI_API_KEY}\n      # 使用默认值：${OPENAI_API_KEY:-sk-default}\n      # 从特定文件加载：${file:.env.production:API_KEY}\n```\n\n完整参考请参阅 [配置指南](docs\u002Fwebsite\u002Fdocs\u002Fconfiguration\u002Findex.md)。\n\n---\n\n## REST API\n\nLlamaFarm 提供与 OpenAI 兼容的 REST API：\n\n**聊天完成**\n```bash\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fprojects\u002Fdefault\u002Fmy-project\u002Fchat\u002Fcompletions \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\n    \"messages\": [{\"role\": \"user\", \"content\": \"你好\"}],\n    \"stream\": false,\n    \"rag_enabled\": true\n  }'\n```\n\n**RAG 查询**\n```bash\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fprojects\u002Fdefault\u002Fmy-project\u002Frag\u002Fquery \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\n    \"query\": \"需求是什么？\",\n    \"database\": \"main_db\",\n    \"top_k\": 5\n  }'\n```\n\n所有端点请参阅 [API 参考](docs\u002Fwebsite\u002Fdocs\u002Fapi\u002Findex.md)。\n\n---\n\n## 专业 ML 能力\n\nUniversal Runtime 提供了超越聊天的功能：\n\n### OCR 与文档提取\n\n```bash\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fvision\u002Focr \\\n  -F \"file=@document.pdf\" \\\n  -F \"model=surya\"\n```\n\n### 异常检测\n\nLlamaFarm 支持 PyOD 提供的 12 种以上异常检测算法，涵盖批处理和流式两种模式。\n\n```bash\n# 基于正常数据训练\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fml\u002Fanomaly\u002Ffit \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"model\": \"sensor-detector\", \"backend\": \"ecod\", \"data\": [[22.1], [23.5], ...]}'\n\n# 检测异常\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fml\u002Fanomaly\u002Fdetect \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"model\": \"sensor-detector\", \"data\": [[22.0], [100.0], [23.0]], \"threshold\": 0.5}'\n\n# 流式检测（处理冷启动、自动重新训练、滑动窗口）\ncurl -X POST http:\u002F\u002Flocalhost:14345\u002Fv1\u002Fml\u002Fanomaly\u002Fstream \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"model\": \"live-sensor\", \"data\": {\"temperature\": 72.5}, \"backend\": \"ecod\"}'\n```\n\n**可用后端：** `ecod`（推荐）、`isolation_forest`、`one_class_svm`、`local_outlier_factor`、`autoencoder`、`hbos`、`copod`、`knn`、`mcd`、`cblof`、`suod`、`loda`\n\n### 文本分类与 NER\n\n完整文档请参阅 [模型指南](docs\u002Fwebsite\u002Fdocs\u002Fmodels\u002Findex.md)。\n\n### 工具调用（MCP）\n\n通过模型上下文协议（MCP）使模型能够访问外部工具：\n\n```yaml\n# 在 llamafarm.yaml 中\nmcp:\n  servers:\n    - name: filesystem\n      transport: stdio\n      command: npx\n      args: ['-y', '@modelcontextprotocol\u002Fserver-filesystem', '\u002Fdata']\n\nruntime:\n  models:\n    - name: assistant\n      provider: ollama\n      model: llama3.1:8b\n      mcp_servers: [filesystem]\n```\n\nLlamaFarm 还将其自身 API 作为 MCP 工具公开，供 Claude Desktop、Cursor 等 MCP 客户端使用。详细信息请参阅 [工具调用指南](docs\u002Fwebsite\u002Fdocs\u002Fmcp\u002Findex.md)。\n\n---\n\n## 示例\n\n| 示例 | 描述 | 位置 |\n|------|------|------|\n| **RAG 示例** | | |\n| 大型复杂 PDF | 数百万字的规划条例 | `examples\u002Flarge_complex_rag\u002F` |\n| 多个小文件 | FDA 来往信函 | `examples\u002Fmany_small_file_rag\u002F` |\n| 混合格式 | PDF、Markdown、HTML、文本和代码 | `examples\u002Fmixed_format_rag\u002F` |\n| 快速笔记 | 小文件快速烟雾测试 | `examples\u002Fquick_rag\u002F` |\n| **异常检测** | | |\n| 快速入门 | 最简单的异常检测示例 | `examples\u002Fanomaly\u002F01_quick_start.py` |\n| 欺诈检测 | 训练、保存、加载模型 | `examples\u002Fanomaly\u002F02_fraud_detection.py` |\n| 流式传感器 | 物联网监控与滚动特征 | `examples\u002Fanomaly\u002F03_streaming_sensors.py` |\n| 后端比较 | 比较所有 12 种算法 | `examples\u002Fanomaly\u002F04_backend_comparison.py` |\n| **用例** | | |\n| FDA 信函助手 | 法规文件分析 | `examples\u002Ffda_rag\u002F` |\n| 政府规划 | 大型条例文件 | `examples\u002Fgov_rag\u002F` |\n\n设置说明和完整列表请参阅 [`examples\u002FREADME.md`](examples\u002FREADME.md)。\n\n---\n\n## 行业用例\n\nLlamaFarm 广泛应用于各行业，用于文档分析、监控和欺诈检测：\n\n- **[制药与治疗](docs\u002Fwebsite\u002Fdocs\u002Fuse-cases\u002Fpharmaceutical-fda.md)** — 分析 FDA 来函，跟踪监管问题\n- **[物联网传感器监控](docs\u002Fwebsite\u002Fdocs\u002Fuse-cases\u002Fiot-sensor-monitoring.md)** — 实时流式异常检测，并支持自动重新训练\n- **[金融欺诈检测](docs\u002Fwebsite\u002Fdocs\u002Fuse-cases\u002Ffinancial-fraud-detection.md)** — 基于速度和行为模式的多阶段欺诈检测\n\n---\n\n## 开发与测试\n\n```bash\n# Python 服务器测试\ncd server && uv sync && uv run --group test python -m pytest\n\n# CLI 测试\ncd cli && go test .\u002F...\n\n# RAG 测试\ncd rag && uv sync && uv run pytest tests\u002F\n\n# 通用运行时测试\ncd runtimes\u002Funiversal && uv sync && uv run pytest tests\u002F\n\n# 构建文档\nnx build docs\n```\n\n---\n\n## 可扩展性\n\n- **添加运行时**：通过实现提供商支持并更新架构来完成\n- **添加向量存储**：通过实现存储后端（Chroma、Qdrant 等）来完成\n- **添加解析器**：用于处理新的文件格式（PDF、DOCX、HTML、CSV 等）\n- **添加提取器**：用于自定义元数据提取\n- **添加 CLI 命令**：在 `cli\u002Fcmd\u002F` 目录下添加\n\n有关详细步骤，请参阅[扩展指南](docs\u002Fwebsite\u002Fdocs\u002Fextending\u002Findex.md)。\n\n---\n\n## 社区与支持\n\n- [Discord](https:\u002F\u002Fdiscord.gg\u002FRrAUXTCVNF) — 与团队及社区交流\n- [GitHub Issues](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues) — 用于报告 bug 和提出功能需求\n- [Discussions](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fdiscussions) — 用于分享想法和提案\n- [贡献指南](CONTRIBUTING.md) — 关于代码风格和贡献流程的说明\n\n---\n\n## 许可证\n\n采用 [Apache 2.0 许可证](LICENSE)授权。感谢名单请参见 [CREDITS](CREDITS.md)。\n\n---\n\n本地构建，随处部署，掌控您的 AI。","# LlamaFarm 快速上手指南\n\nLlamaFarm 是一个完全在本地运行的开源 AI 平台，支持构建 RAG 应用、训练自定义分类器、异常检测及文档处理。数据无需上传云端，确保隐私安全，且支持离线运行。\n\n## 环境准备\n\n### 系统要求\n- **操作系统**：macOS (Universal), Windows, Linux (x86_64 \u002F ARM64)\n- **硬件加速**：自动支持 Apple Silicon, NVIDIA GPU, AMD GPU\n- **内存建议**：运行大模型建议 16GB+ 内存（小模型可在 8GB 运行）\n\n### 前置依赖（仅 CLI\u002F源码模式需要）\n若使用桌面版可跳过此步。若使用命令行或开发模式，请确保安装：\n- **Python**: 3.10+\n- **Go**: 1.24+ (部分底层组件需要)\n- **Node.js & npm**: 用于源码构建\n- **Nx**: 全局安装 (`npm install -g nx`)\n\n> **国内加速提示**：\n> - Python 包安装建议使用清华源：`pip install -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple \u003Cpackage>`\n> - HuggingFace 模型下载若受阻，可配置 `HF_ENDPOINT=https:\u002F\u002Fhf-mirror.com` 环境变量。\n\n---\n\n## 安装步骤\n\n推荐优先使用**桌面应用**，零配置即可启动。开发者可选择 **CLI** 或 **源码** 模式。\n\n### 方式一：桌面应用（推荐新手）\n直接下载对应系统的安装包并运行，无需命令行操作：\n- **Mac**: [下载 DMG](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-mac-universal.dmg)\n- **Windows**: [下载 EXE](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-windows.exe)\n- **Linux**: [下载 AppImage](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Freleases\u002Flatest\u002Fdownload\u002FLlamaFarm-desktop-app-linux-x86_64.AppImage)\n\n### 方式二：命令行工具 (CLI)\n适合集成到工作流或无图形界面服务器。\n\n**macOS \u002F Linux:**\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fllama-farm\u002Fllamafarm\u002Fmain\u002Finstall.sh | bash\n```\n\n**Windows (PowerShell):**\n```powershell\nirm https:\u002F\u002Fraw.githubusercontent.com\u002Fllama-farm\u002Fllamafarm\u002Fmain\u002Finstall.ps1 | iex\n```\n\n### 方式三：源码开发模式\n适合贡献代码或深度定制。\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm.git\ncd llamafarm\n\n# 初始化 Nx 工作区\nnpm install -g nx\nnx init --useDotNxInstallation --interactive=false\n\n# 启动服务（需在三个独立终端分别运行）\nnx start server           # FastAPI 服务 (端口 14345)\nnx start rag              # RAG 文档处理 worker\nnx start universal-runtime # ML 推理引擎 (端口 11540)\n```\n\n---\n\n## 基本使用\n\n以下以 **CLI 模式** 为例，演示如何初始化项目、上传文档并进行问答。\n\n### 1. 初始化并启动项目\n创建新项目并启动本地服务（会自动打开 Designer Web UI）：\n```bash\nlf init my-project\nlf start\n```\n启动后，可在浏览器访问 `http:\u002F\u002Flocalhost:14345` 使用可视化界面。\n\n### 2. 创建数据集并上传文档\n构建一个名为 `research` 的数据集，关联默认处理策略和数据库，并上传 PDF 文件（默认自动解析）：\n```bash\n# 创建数据集\nlf datasets create -s default -b main_db research\n\n# 上传文件（自动进行 OCR 和分块处理）\nlf datasets upload research .\u002Fdocs\u002F*.pdf\n```\n\n### 3. 进行 RAG 问答\n对已上传的文档库进行语义搜索和问答：\n```bash\nlf rag query --database main_db \"这篇文档的核心发现是什么？\"\n```\n\n### 4. 通用聊天测试\n不依赖文档，直接与本地模型对话：\n```bash\n# 交互式对话\nlf chat\n\n# 单次消息\nlf chat \"你好，LlamaFarm！\"\n\n# 指定特定模型对话\nlf chat --model powerful \"解释一下量子纠缠\"\n```\n\n### 5. 配置多模型（可选）\n编辑项目根目录下的 `llamafarm.yaml` 可切换模型提供商（如 Ollama, OpenAI, 或本地 Universal Runtime）：\n\n```yaml\nruntime:\n  models:\n    default:\n      provider: universal\n      model: Qwen\u002FQwen2.5-1.5B-Instruct\n      base_url: http:\u002F\u002F127.0.0.1:11540\u002Fv1\n```\n\n修改配置后，重启服务即可生效。","某金融合规团队需要在完全隔离的内网环境中，快速构建一个能自动解析每日数百份 PDF 财报并提取关键风险指标的本地化 AI 系统。\n\n### 没有 llamafarm 时\n- **数据泄露风险高**：处理敏感财务数据必须依赖云端 API，违反公司“数据不出域”的严格合规要求。\n- **部署门槛极高**：需要手动配置复杂的 Python 环境、GPU 驱动及向量数据库，耗时数天且容易因依赖冲突失败。\n- **运营成本不可控**：按 Token 计费的商业模型导致处理大量文档时成本激增，且无法预测月度预算。\n- **离线能力缺失**：一旦网络波动或中断，整个文档分析流程立即瘫痪，无法保证业务连续性。\n\n### 使用 llamafarm 后\n- **实现绝对隐私保护**：llamafarm 直接在本地硬件运行，所有财报数据全程不离机，完美满足内网合规审计。\n- **分钟级极速上线**：通过桌面应用一键安装，自动调用本地 NVIDIA 或 Apple Silicon 算力，无需编写任何配置代码即可启动 RAG 流水线。\n- **零边际使用成本**：利用开源模型本地推理，彻底消除按量付费开销，无论处理多少文档都无需额外支出。\n- **稳定离线运行**：模型下载后即可在纯离线环境下工作，确保持续、稳定的文档结构化提取与异常检测服务。\n\nllamafarm 让企业能够在自有硬件上以零成本、零风险的方式，几分钟内将复杂的 AI 文档处理能力转化为落地的生产力。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fllama-farm_llamafarm_0a01447d.png","llama-farm","Llama Farm","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fllama-farm_9dc19954.png","Local models, agents, and databases delivered anywhere. ",null,"opensource@llamafarm.dev","https:\u002F\u002Fllamafarm.dev","https:\u002F\u002Fgithub.com\u002Fllama-farm",[84,88,92,96,100,104,108,111,115,118],{"name":85,"color":86,"percentage":87},"Python","#3572A5",59,{"name":89,"color":90,"percentage":91},"TypeScript","#3178c6",30.1,{"name":93,"color":94,"percentage":95},"Go","#00ADD8",8.2,{"name":97,"color":98,"percentage":99},"Shell","#89e051",2.4,{"name":101,"color":102,"percentage":103},"JavaScript","#f1e05a",0.2,{"name":105,"color":106,"percentage":107},"Dockerfile","#384d54",0.1,{"name":109,"color":110,"percentage":107},"HTML","#e34c26",{"name":112,"color":113,"percentage":114},"CSS","#663399",0,{"name":116,"color":117,"percentage":114},"Makefile","#427819",{"name":119,"color":120,"percentage":114},"Batchfile","#C1F12E",830,49,"2026-04-03T19:43:14","Apache-2.0","Linux, macOS, Windows","非必需，但支持自动加速。兼容 Apple Silicon (NPU\u002FGPU)、NVIDIA GPU 和 AMD GPU。未指定具体显存大小或 CUDA 版本要求。","未说明",{"notes":129,"python":130,"dependencies":131},"该工具提供桌面应用程序（无需命令行）和 CLI\u002F源码开发模式。核心架构包含 Server、RAG Worker 和 Universal Runtime 三个服务。支持离线运行（模型下载后）。配置主要通过 llamafarm.yaml 管理。开发模式需安装 Go 1.24+ 和 Nx。","3.10+",[132,133,134,135,136,137,138,139],"FastAPI","Celery","Nx","Go 1.24+","SetFit","PyOD","LlamaIndex","sentence-transformers",[60,14,13,15,26],[142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160],"ai","edge","edge-computing","llama3","llama4","models","aiproject","finetuning-llms","prompt-engineering","rag","chatgpt","claude","gemma","grok","mistral","mlops","openai","qwen","sora","2026-03-27T02:49:30.150509","2026-04-06T12:11:04.493660",[164,169,174,179,184,189],{"id":165,"question_zh":166,"answer_zh":167,"source_url":168},18815,"RAG 系统突然停止工作并报告“已忘记”处理过的文件，甚至出现分段错误（Segfault），该如何解决？","这通常是由于旧版本中的稳定性问题导致的。请尝试升级到 LlamaFarm 的最新版本。用户反馈在更新到最新版本后，该问题不再出现，RAG 系统能正常识别已处理的文件且健康检查通过。","https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F515",{"id":170,"question_zh":171,"answer_zh":172,"source_url":173},18816,"运行 `lf dev` 时服务器健康状态显示为“降级（degraded）”，特别是涉及 Celery 部分的警告，如何处理？","这是一个已知的问题，实际上并非真正的故障。维护者建议忽略或直接移除对 Celery 的状态检查。该警告不影响核心功能，后续版本中可能会修复或移除此检查以避免误导用户。","https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F185",{"id":175,"question_zh":176,"answer_zh":177,"source_url":178},18817,"LlamaFarm 目前支持哪些文件格式的解析和数据提取？","除了基础的 PDF 和 JSON 格式外，项目已扩展支持多种格式。关键解析器包括 Markdown、DOCX、纯文本 (Text)、HTML 和 CSV。此外，还实现了提取器以在不使用 LLM 的情况下提取关键词、实体、摘要、统计信息和元数据，从而增强搜索和上下文能力。","https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F24",{"id":180,"question_zh":181,"answer_zh":182,"source_url":183},18818,"在 macOS 上安装桌面应用后，创建演示项目时在“处理数据集”阶段失败并报错，怎么办？","此问题已在 v0.0.26 及更高版本中修复。如果您遇到此错误，请下载并安装最新版本的 macOS 应用程序。更新后，演示项目（如 Llama 和 Alpaca 百科全书）应能正常创建和运行。","https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F728",{"id":185,"question_zh":186,"answer_zh":187,"source_url":188},18819,"如何查看本地运行服务的最新日志进行调试？","由于项目架构已迁移，不再依赖 Docker 容器，因此无法通过容器命令查看日志。您可以直接查看本地服务生成的日志文件。日志通常存储在用户主目录下的 `.llama` 文件夹中（例如 `~\u002F.llama`），您可以使用文本编辑器或 `cat`\u002F`tail` 命令直接读取这些文件内容。","https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F291",{"id":190,"question_zh":191,"answer_zh":192,"source_url":193},18820,"Designer 界面中的数据库和检索策略配置是否已完全启用？","是的，Designer 中的数据库连接、嵌入策略以及检索策略的输入输出均已连接并正常工作。系统会正确读写配置文件。部分尚未支持的功能组件会在界面中被隐藏或移除，以确保用户体验的流畅性。","https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F359",[195,200,205,210,215,220,225,230,235,240,245,250,255,260,265,270,275,280,285,290],{"id":196,"version":197,"summary_zh":198,"released_at":199},109340,"v0.0.28","## [0.0.28](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.27...v0.0.28) (2026-03-05)\n\n\n### 功能特性\n\n* **cli:** 添加部署命令和捆绑包打包功能 ([#772](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F772)) ([da1bb13](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fda1bb13b8429bd044712f7f8aed023cbac655e5f))\n* **designer:** 部署工作流的捆绑包 UI ([#783](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F783)) ([906b9e3](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F906b9e31c6be48acc330ef2156616032e7d11169))\n* 机器学习插件 — 时间序列、ADTK、漂移检测、CatBoost（基于插件，可按需选择） ([#766](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F766)) ([55850d6](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F55850d6dae7c28450d5fc15b1e9c12f6ee434c97))\n* **runtime:** 为 GGUF 聊天补全添加 logprobs 支持 ([#791](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F791)) ([79c9874](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F79c98741b031d1dd138e38bba3a8b88017bb8642))\n* 带有多轮链式调用和预热功能的服务器端 KV 缓存 ([#782](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F782)) ([e750ed1](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe750ed19b5d25d83ca2c66173719844533576d9f))\n* **server:** 添加内置工具系统及任务工具 ([#738](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F738)) ([952597c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F952597c51a33799b64001ded3dd3b1c8196ede8b))\n* 支持结构化输出 ([#724](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F724)) ([55fa703](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F55fa703702cc18a167fc6521a8bfd1f548e3ba4e))\n* 视觉 API 改进、评估流水线以及目标跟踪 ([#780](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F780)) ([f9635a4](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff9635a476e118cec8f79bd49ae00c5c1d3feb1f5))\n* 设计师端的视觉 UI ([#777](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F777)) ([6542c74](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F6542c7474a6d8a65af7790b0edee23992893e5e6))\n* **vision:** 检测+分类组合、会话 TTL 清理、COCO 数据集支持 ([#774](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F774)) ([96151ab](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F96151ab2ae3d9cf7b0634b4109d7a30635d29984))\n* **vision:** 视觉 MVP — 检测、分类、训练、级联、反馈 ([#765](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F765)) ([bda3c6e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fbda3c6e1d31444e4ba20c64ffa4b0f90abb1be44))\n\n\n### 错误修复\n\n* **ci:** 从文件中读取 prose 变更日志以避免 ARG_MAX 溢出 ([ccae0a0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fccae0a0c4402ddd45ef782d317b909d2b971abf3))\n* **cli:** 将插件注册表嵌入到发布版的二进制文件中 ([#792](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F792)) ([38b481e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F38b481e5219c0ec56cefadd116b08d4b7fa42dcc))\n* 内容预算计算 ([#779](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafa","2026-03-05T19:42:42",{"id":201,"version":202,"summary_zh":203,"released_at":204},109341,"v0.0.27","## [0.0.27](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.26...v0.0.27) (2026-02-16)\n\n\n### 功能特性\n\n* 插件 ([#748](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F748)) ([c026511](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fc026511bbb2d5a58ba7d514cdc32ad04e3af0411))\n* **异常检测：** 添加全面的文档、用例和完整演示 ([#715](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F715)) ([fe988ad](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ffe988ad572e0e6542c2b2aba4392d686e433f2cc))\n* 二进制组件构建 ([#681](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F681)) ([a9858e4](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fa9858e48546d668386826dced2a0cf897d0aa7c4))\n* **配置：** 添加按模型的 RAG 默认设置 ([#759](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F759)) ([4ad4a6c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F4ad4a6cd40916b383e5123c7d91798757b6e07db))\n* **设计器：** 在测试输出中显示 RAG 来源片段 ([#740](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F740)) ([be5ea97](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fbe5ea9719fb277d6bde364ee44fed4703a1cca07))\n* **服务器：** 级联默认的数据处理策略 ([#697](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F697)) ([882a643](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F882a643f00a5842a58579a9cda3921f4472763ac))\n* **服务器：** 将默认端口从 8000 更改为 14345 ([#731](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F731)) ([7150e75](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F7150e755c6e538e75a17b28f4dfd050eaffd4069))\n\n\n### 错误修复\n\n* **音频：** 设计器错误处理  ([#722](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F722)) ([dc41a42](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fdc41a424ce054662adc235fee0801810289b6d57))\n* **设计器：** 改善删除用户体验并处理遗留项目 ([#739](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F739)) ([cb002d4](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fcb002d43d44ece7df5e6559a8c29e71f7eca5661))\n* **设计器：** 打开删除模态时保留项目名称 ([10ba578](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F10ba578deb043de33c23356f3aff820c5b8e76ae))\n* **设计器：** 移除导致训练按钮出现 404 错误的冗余保存调用 ([#741](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F741)) ([05d0779](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F05d0779ba9310fb8e92f65f9595d1f0198ea7caf))\n* **设计器：** 按顺序安装插件，并在安装后自动启用 ([#767](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F767)) ([c4e78f9](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fc4e78f9701f39543bb1f38423a05126ea4c10ee1))\n* **设计器：** 在演示项目转换后更新入门检查清单 ([#745](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F745)) ([1e8ee89](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1e8ee899ba9dff7f39c53b87ec9d594f73cc694b))\n* **运行时：** 添加智能 GPU 分配功能，以防止多模型 OOM 崩溃 ([#761](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F761)) ([848d10","2026-02-16T21:28:30",{"id":206,"version":207,"summary_zh":208,"released_at":209},109342,"v0.0.1+feat.addons","用于测试 feat-addons 分支的预发布版本。","2026-02-10T15:52:28",{"id":211,"version":212,"summary_zh":213,"released_at":214},109343,"v0.0.26","## [0.0.26](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.25...v0.0.26) (2026-01-27)\n\n\n### 功能特性\n\n* **config:** 添加可复用组件 ([#682](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F682)) ([e59ec87](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe59ec871a935285ff2bff27000beda90c1e6f3d1))\n* **designer:** 添加园艺和家居维修示例数据集 ([#688](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F688)) ([b2c71f3](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fb2c71f34b5cb2fd735dfb97ab9abb2472d0c9889))\n* **designer:** 添加端口并清理链接 ([2d03d2c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F2d03d2cce0c2878e1ac6baf44b89490a0f852355))\n* **designer:** 在页眉中添加服务状态面板 ([3d9aca7](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F3d9aca77694c4d7cbc9deacf4d1076e01227d950))\n* **designer:** 自动处理数据集 ([#679](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F679)) ([b15db41](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fb15db413c61ea9e2a2c48c21b55bf96e277af14a))\n* **designer:** 小幅更新颜色、尺寸和布局 ([#711](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F711)) ([8bd9542](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F8bd95426478d8f3866a90849a75c574a238ce981))\n* **electron-app:** 改进启动画面用户体验及跨平台支持 ([#712](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F712)) ([35f56f7](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F35f56f70ca986ffaaf7d8c02e0b8460c6c65729b))\n* 全双工语音推理流水线 ([#690](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F690)) ([0d6090b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0d6090baa2ed3342a981cc5713943970944cc9e6))\n* **rag:** 添加带策略选择的文档预览 ([#704](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F704)) ([87fe04f](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F87fe04fab19920fdc45cd29744bb58fec201f5ca))\n* **rag:** 通用RAG——零配置默认策略 ([#696](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F696)) ([909f549](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F909f549d3713035342715a7fc375d09dd8e59b3b))\n* **runtime:** 音频处理 ([#685](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F685)) ([3d7536c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F3d7536cda9e4872aefdd9b867aa342af14dd61ea))\n* **server:** 为提示词和工具添加动态值替换功能 ([#672](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F672)) ([127fccb](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F127fccb51b05471705248c3cacd8801a6f0705f7))\n\n\n### 错误修复\n\n* **cli:** 防止开发构建停止正在运行的服务 ([#699](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F699)) ([2173f77](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F2173f773ff36ebe0a6dcea08e0229da11d5f6758)), 关闭了 [#698](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F698)\n* **cli:** 移除 PYTORCH_MPS_HIGH_WATERMARK_RATIO 以修复示例项目创建问题 ([#725](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F725)) ([1fdfec7](https","2026-01-27T21:45:54",{"id":216,"version":217,"summary_zh":218,"released_at":219},109344,"v0.0.25","## [0.0.25](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.24...v0.0.25) (2026-01-14)\n\n\n### 功能特性\n\n* 数据集上传时自动处理文件 ([#661](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F661)) ([e08d762](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe08d762d73baf128023ab571352c88b8e88a1533))\n* **设计器:** 将所有 API 调用添加到开发者工具中 ([#680](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F680)) ([d36f6ff](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fd36f6ffb9967d9e43027c7761c2395d5ca94b4ee))\n* **设计器:** 为嵌入模型下载添加 SSE 流式传输功能 ([#655](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F655)) ([b2ffbd0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fb2ffbd0319f6bab7c165fd78a1c87cbfc9738833))\n* **设计器:** 测试空间，包含异常检测和分类器测试以及其他更新 ([#657](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F657)) ([fc94563](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ffc94563cbcde16614dddb6bcddc4a9c0fe93077d))\n* **运行时:** 支持原生工具调用 ([#636](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F636)) ([f99f305](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff99f30538ffc4d42be46ba1cfd43ef1cc758f935))\n\n\n### 错误修复\n\n* **CLI:** 配置验证错误输出 ([#673](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F673)) ([6d92f43](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F6d92f43913cc5cbd135516a624b20900f4c55970))\n* 在配备 NVIDIA 显卡的 Windows 系统上安装\u002F运行失败问题 ([#665](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F665)) ([29f3242](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F29f324246040d1db6cc45faf570fdd60d0396cce))\n* **RAG:** 移除解析器回退机制 ([#642](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F642)) ([7037219](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F7037219de690a4518bf5acb16afe730d390faa82))\n* **运行时:** 将依赖项移至主模块，并启用离线 GGUF 加载功能 ([#670](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F670)) ([1d9de0c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1d9de0ce9cc80c1ad0bdf4bdbbfef8af75c3c21d))\n\n\n### 其他杂项任务\n\n* 发布版本 0.0.25 ([74dea58](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F74dea58befffae34d9a3ac1072c680b3e20b5ea5))\n\n---\n此 PR 由 [Release Please](https:\u002F\u002Fgithub.com\u002Fgoogleapis\u002Frelease-please) 生成。请参阅 [文档](https:\u002F\u002Fgithub.com\u002Fgoogleapis\u002Frelease-please#release-please)。\n\n\u003C!-- CURSOR_SUMMARY -->\n\n## 桌面应用下载\n\n### macOS\n适用于 Apple Silicon (M1+) 或 Intel Mac 的下载：\n- **Apple Silicon (arm64)**: `LlamaFarm-desktop-app-mac-arm64.dmg` 或 `.zip`\n- **Intel (x64)**: `LlamaFarm-desktop-app-mac-x64.dmg` 或 `.zip`\n\n**安装步骤：**\n1. 下载适合您架构的 DMG 或 ZIP 文件。\n2. 打开 DMG 并将 LlamaFarm 拖拽到“应用程序”文件夹（或解压 ZIP 文件）。\n3. 启动应用（已签名并经过公证，因此可正常打开）。\n4. 应用会自动安装 CLI 并启动相关服务。\n\n### Linux\n下载适合您 Linux 发行版的相应软件包：\n- **AppImage**（通用）：`","2026-01-14T06:11:18",{"id":221,"version":222,"summary_zh":223,"released_at":224},109345,"v0.0.24","## [0.0.24](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.23...v0.0.24) (2026-01-06)\n\n\n### 功能特性\n\n* **anomaly:** 为异常检测分数添加归一化方法 ([#624](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F624)) ([9850015](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F9850015b19696374fabdb1c8bbaf2e43b73eb24d))\n* **designer:** 异常检测的用户体验优化 ([#635](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F635)) ([2754881](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F27548810b35a813797ce5b66ae8da52eff26b9bf))\n\n\n### Bug修复\n\n* **designer:** 异常检测和分类器的用户体验修复  ([#645](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F645)) ([5610c35](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F5610c35a2e75adc8d6669914abeb7bf97dbae84a))\n\n\n### 其他杂项\n\n* 发布版本 0.0.24 ([0501ed7](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0501ed71c12224bc6f27f81e931c838c47249833))\n\n## 桌面应用下载\n\n### macOS\n适用于 Apple Silicon (M1+) 或 Intel Mac 的下载：\n- **Apple Silicon (arm64)**: `LlamaFarm-desktop-app-mac-arm64.dmg` 或 `.zip`\n- **Intel (x64)**: `LlamaFarm-desktop-app-mac-x64.dmg` 或 `.zip`\n\n**安装步骤：**\n1. 下载适合您架构的 DMG 或 ZIP 文件。\n2. 打开 DMG 并将 LlamaFarm 拖到“应用程序”文件夹中（或解压 ZIP 文件）。\n3. 启动应用（已签名并经过公证，因此可正常打开）。\n4. 应用会自动安装 CLI 并启动相关服务。\n\n### Linux\n根据您的 Linux 发行版下载相应软件包：\n- **AppImage**（通用）: `LlamaFarm-desktop-app-linux.AppImage`\n- **Debian\u002FUbuntu**: `LlamaFarm-desktop-app-linux.deb`\n\n**安装步骤：**\n```bash\n# AppImage（适用于所有发行版）\nchmod +x LlamaFarm-desktop-app-linux.AppImage\n.\u002FLlamaFarm-desktop-app-linux.AppImage\n\n# Debian\u002FUbuntu\nsudo dpkg -i LlamaFarm-desktop-app-linux.deb\n```\n\n### Windows\n下载 Windows 安装程序：\n- **Windows Installer**: `LlamaFarm-desktop-app-windows.exe`\n\n**安装步骤：**\n1. 下载安装程序。\n2. 运行安装程序（会安装到您的用户目录）。\n3. 从开始菜单启动 LlamaFarm。\n4. 应用会自动安装 CLI 并启动相关服务。\n\n---\n\n### 自动更新\n安装完成后，应用会自动检查更新，并提示您进行安装。","2026-01-06T21:05:52",{"id":226,"version":227,"summary_zh":228,"released_at":229},109346,"v0.0.23","## [0.0.23](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.22...v0.0.23) (2025-12-20)\n\n\n### Bug修复\n\n* **运行时:** 由于日志记录问题导致的管道破裂 ([02969d0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F02969d0103cbdb8c7a30c0abd0ac7881a5ea33b9))\n\n## 桌面应用下载\n\n### macOS\n适用于 Apple Silicon (M1+) 或 Intel Mac 的下载：\n- **Apple Silicon (arm64)**: `LlamaFarm-desktop-app-mac-arm64.dmg` 或 `.zip`\n- **Intel (x64)**: `LlamaFarm-desktop-app-mac-x64.dmg` 或 `.zip`\n\n**安装步骤：**\n1. 下载适合您设备架构的 DMG 或 ZIP 文件。\n2. 打开 DMG 文件，将 LlamaFarm 拖拽到“应用程序”文件夹中（或解压 ZIP 文件）。\n3. 启动应用（该应用已签名并经过公证，因此可正常打开）。\n4. 应用会自动安装 CLI 并启动相关服务。\n\n### Linux\n根据您的 Linux 发行版下载相应软件包：\n- **AppImage**（通用）: `LlamaFarm-desktop-app-linux.AppImage`\n- **Debian\u002FUbuntu**: `LlamaFarm-desktop-app-linux.deb`\n\n**安装步骤：**\n```bash\n# AppImage（适用于所有发行版）\nchmod +x LlamaFarm-desktop-app-linux.AppImage\n.\u002FLlamaFarm-desktop-app-linux.AppImage\n\n# Debian\u002FUbuntu\nsudo dpkg -i LlamaFarm-desktop-app-linux.deb\n```\n\n### Windows\n下载 Windows 安装程序：\n- **Windows 安装程序**: `LlamaFarm-desktop-app-windows.exe`\n\n**安装步骤：**\n1. 下载安装程序。\n2. 运行安装程序（它会安装到您的用户目录）。\n3. 从“开始”菜单启动 LlamaFarm。\n4. 应用会自动安装 CLI 并启动相关服务。\n\n---\n\n### 自动更新\n应用安装完成后，会自动检查更新，并提示您进行安装。","2025-12-20T06:09:32",{"id":231,"version":232,"summary_zh":233,"released_at":234},109347,"v0.0.22","## [0.0.22](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.21...v0.0.22) (2025-12-19)\n\n\n### 错误修复\n\n* **通用：** 将 logits_processor 作为可调用对象传递，而不是列表 ([#628](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F628)) ([84cf924](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F84cf924e1494877a01c7bcfb01d81447c2746531)), 关闭了 [#627](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F627)\n\n## 桌面应用下载\n\n### macOS\n适用于 Apple Silicon (M1+) 或 Intel Mac 的下载：\n- **Apple Silicon (arm64)**：`LlamaFarm-desktop-app-mac-arm64.dmg` 或 `.zip`\n- **Intel (x64)**：`LlamaFarm-desktop-app-mac-x64.dmg` 或 `.zip`\n\n**安装步骤：**\n1. 下载适合您硬件架构的 DMG 或 ZIP 文件。\n2. 打开 DMG 文件，将 LlamaFarm 拖拽到“应用程序”文件夹中（或解压 ZIP 文件）。\n3. 启动应用（该应用已签名并经过公证，因此可正常打开）。\n4. 应用会自动安装 CLI 并启动相关服务。\n\n### Linux\n根据您的 Linux 发行版下载相应软件包：\n- **AppImage**（通用）：`LlamaFarm-desktop-app-linux.AppImage`\n- **Debian\u002FUbuntu**：`LlamaFarm-desktop-app-linux.deb`\n\n**安装步骤：**\n```bash\n# AppImage（适用于所有发行版）\nchmod +x LlamaFarm-desktop-app-linux.AppImage\n.\u002FLlamaFarm-desktop-app-linux.AppImage\n\n# Debian\u002FUbuntu\nsudo dpkg -i LlamaFarm-desktop-app-linux.deb\n```\n\n### Windows\n下载 Windows 安装程序：\n- **Windows 安装程序**：`LlamaFarm-desktop-app-windows.exe`\n\n**安装步骤：**\n1. 下载安装程序。\n2. 运行安装程序（它会安装到您的用户目录）。\n3. 从开始菜单启动 LlamaFarm。\n4. 应用会自动安装 CLI 并启动相关服务。\n\n---\n\n### 自动更新\n应用安装完成后，会自动检查更新，并提示您进行安装。","2025-12-19T22:02:31",{"id":236,"version":237,"summary_zh":238,"released_at":239},109348,"v0.0.21","## [0.0.21](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.20...v0.0.21) (2025-12-19)\n\n\n### 功能特性\n\n* **api:** 添加视觉路由和 ML 端点的模型版本控制（[#608](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F608)）（[41b93fa](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F41b93fa842fc93b79392e68f5580aec8f01b7566)）\n* **designer:** 添加圣诞老人假日助手演示（[#618](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F618)）（[0df6f1e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0df6f1e7f5d6091111438d2cca03261e96570237)）\n* **designer:** 数据增强（[#591](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F591)）（[42b49e5](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F42b49e5b4517730607855cbc488a20675ac9b9e3)）\n* **designer:** 优化 RAG 用户体验（[a43e798](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fa43e7981e4fe8f7388e62035f864765478fd334d)）\n* **designer:** 增加检索策略设置以测试聊天功能（[#597](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F597)）（[f6bd7a0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff6bd7a097cef47823ff2a4fe0ec0eda17a502ca7)）\n* **designer:** 测试聊天功能的修复与改进（[#609](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F609)）（[47a25af](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F47a25af7076c355155d2a3dce2502df3588edfc6)）\n* 全局项目列表命令（[#604](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F604)）（[30a9185](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F30a9185688bd6b0042e756820f82f81888eb120d)）\n* **runtime:** 包含适用于所有平台的原生 llama-cpp 绑定（[#603](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F603)）（[df70282](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fdf70282ba7a3c5e70ba9bcaa2a55fe08add6f725)）\n* **universal:** 添加专用 ML 模型——OCR、文档提取、异常检测（[#532](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F532)）（[838eafd](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F838eafde07b2bab1d628770823b9241c5284f6c9)）\n\n\n### 错误修复\n\n* **app:** 确保 Windows 上的 lf 二进制文件带有 .exe 后缀（[#596](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F596)）（[e9914c9](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe9914c91d60d7050ce51dd8791d9b5f80a149845)）\n* **app:** 确保多架构 Linux 构建可用（[#586](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F586)）（[1931098](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F19310986f406302989188db0b8e2c4c889f8c520)）\n* **cli:** 由于无效的复制操作，Linux 上的升级失败（[#585](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F585)）（[fcd3877](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ffcd3877ab0d8f9f4fb7a576bf6fafc26c2cb3a6b)）\n* **designer:** 在首页添加仓库星标标签，并修复提示消息显示问题（[#601](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F601)）（[a13bc4a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fa13bc4a70f79856c6b6f3b352fac3ea30ed8a1e7)）\n* **designer:** 移除控制台日志输出（[28d8291](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F28d8291b80979e2299db770e2788182472ee8e7f)）\n* 模型卸载清理","2025-12-19T21:02:20",{"id":241,"version":242,"summary_zh":243,"released_at":244},109349,"v0.0.20","## [0.0.20](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.19...v0.0.20) (2025-12-10)\n\n\n### 功能特性\n\n* **cli:** 添加自动启动服务标志位 ([#335](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F335)) ([6f18bde](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F6f18bdec1300536dbd1ebe7021aa0fc6b4ef06c8))\n* **designer:** 提供更多可下载的 GGUF 模型选项 ([#546](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F546)) ([7f67d97](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F7f67d97153b21f6a07d1bc9a9f448577daed879a))\n* **rag:** 列出 RAG 数据库中的文档 ([#547](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F547)) ([b99eebc](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fb99eebc55a4172fbe7a954b6ab70026bcef74db8))\n* **rag:** 报告 RAG 统计信息 ([#543](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F543)) ([d6e2837](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fd6e283748d0bf63566fae7d9316f0d7fef311251))\n* 在删除文件时移除数据库中的分块数据 ([#549](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F549)) ([21bd24b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F21bd24b1c6ff7fba8cdbb54428fe835afaa1cac5))\n* **server:** 启动和停止数据处理进程 ([#489](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F489)) ([1448b1a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1448b1ab3cb0a41df9da64c0ff6d897093b1decb))\n\n\n### 错误修复\n\n* **app:** 首次运行时的启动失败问题 ([#567](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F567)) ([f70436e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff70436e12e48aac5f0bf8701fee08ed78493bbb6))\n* **cli:** 在路径解析前展开 ~ 符号 ([#563](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F563)) ([a862716](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fa8627165c6e95c87d8662cb55ced0e6928e7c2a3))\n* **cli:** 忽略 Python 特定的环境变量 ([#564](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F564)) ([ee4f9af](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fee4f9af705f5a1e843d44620916e369eeb2ae892))\n* **cli:** 改进进程管理器的锁机制 ([#573](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F573)) ([863a85a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F863a85a2daff56a650ca5e934972e824d60feaa3))\n* **cli:** 解决因进程停止死锁导致的升级卡顿问题 ([#566](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F566)) ([5466de3](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F5466de34bbb0f6981533cece93731e60497cca54))\n* **designer:** 显示错误版本号 ([#552](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F552)) ([c10335f](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fc10335f8fd0357667d49568d6f05cff49eaa09c4))\n* **rag:** 防止存储失败的向量 ([#571](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F571)) ([4aceb2e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F4aceb2e93a177505752863fb1bd9d92c04b811c9))\n* 显示模态错误对话框 ([#570](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F570)) ([70864f1](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F70864f11b23ecef09354595a1b7beda5368754d3))\n\n\n### 其他杂项工作\n\n* 发布","2025-12-10T17:34:57",{"id":246,"version":247,"summary_zh":248,"released_at":249},109350,"v0.0.19","## [0.0.19](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.18...v0.0.19) (2025-12-03)\n\n\n### Features\n\n* Add automatic model download management ([#545](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F545)) ([e362e12](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe362e123d4cf73e8e1737c3c1afe5d5c066637b0))\n* Add custom RAG query support to chat\u002Fcompletions endpoint ([#536](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F536)) ([dd402a0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fdd402a0eea697069b6dd281b774628a96a78645b))\n* Add thinking\u002Freasoning model support to Universal Runtime ([#542](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F542)) ([e80725e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe80725e67bbad93b7d5c799de24e92f90c39796b))\n* **api:** Add Database CRUD API endpoints ([#524](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F524)) ([e2f0e99](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe2f0e99c328e54aa8f703adbba82421872196f62))\n* **designer:** better ux for day 2 users ([#509](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F509)) ([7cbbafe](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F7cbbafea1a0d6ef80c9e79bad318184f49ea169b))\n* **designer:** check disk space for model download ([#527](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F527)) ([555c094](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F555c0947dd052583a31a5f89a2366f86d1a45c40))\n* **designer:** list gguf models for download ([#503](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F503)) ([0afa661](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0afa66148af8a5857f21321b21e79702c3c4e182))\n\n\n### Bug Fixes\n\n* **api:** Support datasets endpoint without trailing slash ([#519](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F519)) ([e3acfe0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe3acfe04ba49f8454082e9b93200fcb9cd26d411)), closes [#518](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F518)\n* **cli:** handle cross-fs data moves ([#537](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F537)) ([c4ad4fa](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fc4ad4fa4cae5317814251dc9caa58cf01a9d7d1b))\n* Datasets CLI file count and RAG PDF parsing ([#530](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F530)) ([ac6f4e4](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fac6f4e4eec9a9c364c7c2a5d8fc15052b2d91cdd))\n* **demo:** change timeout for inital  demo experience ([#539](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F539)) ([a2dcaa0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fa2dcaa04e047b57c46f57c71911636dc653b337f))\n* **designer:** address build error ([0d194ef](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0d194ef69002145467f17cbd60298cc32268b425))\n* **designer:** fix demo data import and update toast component ([#506](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F506)) ([11a375a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F11a375aa16dceccc3d551015dcba46d104c3b802))\n* **rag:** Apply reranking even when query decomposition fails ([#517](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F517)) ([5030cca](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F5030cca89e56c7bda8d9b6aad651943872e913c6))\n\n\n### Miscellaneous Chores\n\n* release 0.0.19 ([ece3f40](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fece3f40f94c94b2b62fb9647bdee53081acadbb9))\n\n## Desktop App Downloads\n\n### macOS\nDownload the universal DMG that works on both Apple Silicon and Intel Macs:\n- **Universal (Apple Silicon and Intel)**: `LlamaFarm-0.0.19.dmg`\n\n**Installation:**\n1. Download the DMG file\n2. Open the DMG and drag LlamaFarm to Applications\n3. Launch the app (it's signed and notarized, so it will open normally)\n4. The app will auto-install the CLI and start services\n\n### Linux\nDownload the appropriate package for your Linux distribution:\n- **AppImage** (Universal): `LlamaFarm-0.0.19.AppImage`\n- **Debian\u002FUbuntu**: `llamafarm-desktop_0.0.19_amd64.deb`\n\n**Installation:**\n```bash\n# AppImage (works on all distributions)\nchmod +x LlamaFarm-0.0.19.AppImage\n.\u002FLlamaFarm-0.0.19.AppImage\n\n# Debian\u002FUbuntu\nsudo dpkg -i llamafarm-desktop_0.0.19_amd64.deb\n```\n\n### Windows\nDownload the Windows installer:\n- **Windows Installer**: `LlamaFarm-Setup-0.0.19.exe`\n\n**Installation:**\n1. Download the installer\n2. Run the installer (it will install to your user directory)\n3. Launch LlamaFarm from the Start Menu\n4. The app will auto-install the CLI and start services\n\n---\n\n### Auto-Updates\nOnce installed, the app will automatically check for updates and prompt you to install them.\n","2025-12-03T19:01:37",{"id":251,"version":252,"summary_zh":253,"released_at":254},109351,"v0.0.18","## [0.0.18](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.17...v0.0.18) (2025-11-25)\n\n\n### Features\n\n* **app:** sign windows and mac apps ([#502](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F502)) ([f020648](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff0206487c86bee4524b2eea6e587fd63b7a43cba))\n* **rag:** Add advanced retrieval strategies with cross-encoder reranking and multi-turn RAG ([#476](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F476)) ([9a18fe7](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F9a18fe7a600dd40c0193445606ace201630e9c9e))\n\n\n### Bug Fixes\n\n* **cli:** ensure logs are on for services ([#505](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F505)) ([81ab388](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F81ab3882f3acab157d131b65df7b88d35bfcd2f2))\n\n\n### Miscellaneous Chores\n\n* release 0.0.18 ([37b7024](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F37b70246e1866f59a3320e3748e0a5b6aecdb921))\n\n## Desktop App Downloads\n\n### macOS\nDownload the universal DMG that works on both Apple Silicon and Intel Macs:\n- **Universal (Apple Silicon and Intel)**: `LlamaFarm-0.0.18.dmg`\n\n**Installation:**\n1. Download the DMG file\n2. Open the DMG and drag LlamaFarm to Applications\n3. Launch the app (it's signed and notarized, so it will open normally)\n4. The app will auto-install the CLI and start services\n\n### Linux\nDownload the appropriate package for your Linux distribution:\n- **AppImage** (Universal): `LlamaFarm-0.0.18.AppImage`\n- **Debian\u002FUbuntu**: `llamafarm-desktop_0.0.18_amd64.deb`\n\n**Installation:**\n```bash\n# AppImage (works on all distributions)\nchmod +x LlamaFarm-0.0.18.AppImage\n.\u002FLlamaFarm-0.0.18.AppImage\n\n# Debian\u002FUbuntu\nsudo dpkg -i llamafarm-desktop_0.0.18_amd64.deb\n```\n\n### Windows\nDownload the Windows installer:\n- **Windows Installer**: `LlamaFarm-Setup-0.0.18.exe`\n\n**Installation:**\n1. Download the installer\n2. Run the installer (it will install to your user directory)\n3. Launch LlamaFarm from the Start Menu\n4. The app will auto-install the CLI and start services\n\n---\n\n### Auto-Updates\nOnce installed, the app will automatically check for updates and prompt you to install them.\n","2025-11-25T19:40:48",{"id":256,"version":257,"summary_zh":258,"released_at":259},109352,"v0.0.17","## [0.0.17](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.16...v0.0.17) (2025-11-24)\n\n\n### Bug Fixes\n\n* **designer:** fix empty prompts array for new projects, fix current version in designer ([#495](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F495)) ([488313c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F488313cd239c2075f2adcce47b297e0b89005494))\n* **docs:** add troubleshooting and update styles ([#498](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F498)) ([270af60](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F270af6050c3007744aa1e8ea778afe6ce3f66177))\n* hf progress bars can crash runtime ([#499](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F499)) ([50d66a8](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F50d66a847062b5071e47b66231e1b9e9297f2e88))","2025-11-24T21:03:13",{"id":261,"version":262,"summary_zh":263,"released_at":264},109353,"v0.0.16","## [0.0.16](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.15...v0.0.16) (2025-11-23)\n\n\n### Bug Fixes\n\n* cli packaging ([8058b1f](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F8058b1f3f8c809e5896a47e3d73a579d8a14fbfd))","2025-11-23T03:52:17",{"id":266,"version":267,"summary_zh":268,"released_at":269},109354,"v0.0.15","## [0.0.15](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.14...v0.0.15) (2025-11-22)\n\n\n### Features\n\n* Add universal event logging for observability ([#383](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F383)) ([e0c4c8e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe0c4c8e89582a8ab117f28f5cd9af3fbffd6080c))\n* create project from existing project ([#459](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F459)) ([3fee20d](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F3fee20d3801953bd69c687c10c28bb82e925b8d2))\n* **designer:** Add interactive demo project creation system ([#473](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F473)) ([4328380](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F4328380e589f5a7db95f94769f1eb41afe65711d))\n* **designer:** demo app, copy config, test fixes ([#477](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F477)) ([c840da7](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fc840da7a29c817baf29ed86bc886b000cf1a466b))\n* **designer:** import demo data, demo to universal runtime, fix empty prompts array bug ([#491](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F491)) ([dc5a33f](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fdc5a33fc368fc8dbd2d29704be35a441d61f5cf1))\n* **electron:** Desktop app with auto-updates and polished UI ([#458](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F458)) ([02ddf58](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F02ddf5875fce359f3df7d6d5be6ca822aafb0b01))\n* implement enhanced tool calling ([#460](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F460)) ([72d9b82](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F72d9b82abee8611d00ee4290556b751736d22992))\n* **runtime:** support gguf models ([#445](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F445)) ([6479372](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F64793727a30651d2de771410dd8f56a1f8604222))\n\n\n### Bug Fixes\n\n* Add GH_TOKEN to desktop app workflows ([#480](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F480)) ([6402da0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F6402da075ec2a7a90743ef6956215a8d188d39f8))\n* **cli:** prevent uv from building source on run ([#482](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F482)) ([082a269](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F082a269673c3342f2bfa535c877d5fcbfb6aeb8b))\n* **cli:** sync newer config on startup instead of always preferring CWD ([#484](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F484)) ([458cd7b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F458cd7bfa48655ec92963bb36444ad52e142a22a))\n* **cli:** upgrades may fail on unix-like environments ([#453](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F453)) ([3efed4c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F3efed4c31c1f684be9aa3e2078aee41907e06ede))\n* **designer:** database tab switching ([#455](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F455)) ([e8be498](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fe8be498a96b1adec475460d0ca1eff980ec33a67))\n* **designer:** dataset name validation, status column, other polishes ([#472](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F472)) ([bcc8669](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fbcc86699c237cc2c95fd1b4a49a1ffc71175e907))\n* **designer:** Improve RAG integration and chat context management ([#471](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F471)) ([7bb1011](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F7bb1011d13ab16148eb5c3a383f03de8ad19fc9e))\n* **mcp:** handle missing mcp config section gracefully ([#485](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F485)) ([517d663](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F517d663b303d315a1ace964669fb7d5d94eed5a4))\n\n\n### Miscellaneous Chores\n\n* release 0.0.15 ([2395ddc](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F2395ddcc68a71ff500dd6a20dbabe69ba2cdac63))","2025-11-22T05:22:46",{"id":271,"version":272,"summary_zh":273,"released_at":274},109355,"v0.0.14","## [0.0.14](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.13...v0.0.14) (2025-11-13)\n\n\n### Features\n\n* **cli:** show size stats in raq query result ([#442](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F442)) ([f602692](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff6026924ba03fdc719c16481d4cd385ad9b754e4))\n* **designer:** database embedding and retrieval strategies ([#438](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F438)) ([db5ee11](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fdb5ee110dd840b00f8cb8d476d087d0594b4a0be))\n\n\n### Bug Fixes\n\n* **designer:** build chat errors ([#444](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F444)) ([964ed95](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F964ed95f2268c787a5dbe01aa9025e0bb3049263))\n* **designer:** file drop dataset selection ([#423](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F423)) ([1bd0a42](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1bd0a429103b3cc8d3ff69b0ff804862ee92b506))\n* **package:** missing ruff config for datamodel gen ([#450](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F450)) ([1e5a31b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1e5a31bad7bec5a2bf4c8ea4932fffe47896bf06))\n\n\n### Miscellaneous Chores\n\n* release 0.0.14 ([58522a1](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F58522a1a681030d1b4328b23da39b06c1acf676d))","2025-11-13T16:13:35",{"id":276,"version":277,"summary_zh":278,"released_at":279},109356,"v0.0.13","## [0.0.13](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.12...v0.0.13) (2025-11-11)\n\n\n### Bug Fixes\n\n* **cli:** use correct version number ([#440](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F440)) ([d01f9b8](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fd01f9b872b20e8e283daafe30a1e313c89682d99))\n\n---\nThis PR was generated with [Release Please](https:\u002F\u002Fgithub.com\u002Fgoogleapis\u002Frelease-please). See [documentation](https:\u002F\u002Fgithub.com\u002Fgoogleapis\u002Frelease-please#release-please).\n\n## Summary by Sourcery\n\nBug Fixes:\n- Correct the CLI version number output\n\n\u003C!-- CURSOR_SUMMARY -->","2025-11-11T17:48:50",{"id":281,"version":282,"summary_zh":283,"released_at":284},109357,"v0.0.12","## [0.0.12](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.11...v0.0.12) (2025-11-11)\n\n\n### Features\n\n* add tests for project deletion ([afdf3a5](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fafdf3a51e4b54313ff1345528009c068468052c4))\n* **cli:** add delete projects command ([#420](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F420)) ([0a03d2a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0a03d2adf54a2d3b3bc09ddacfe3ba1b2a502441))\n* **designer:** add copy button to config editor ([af22097](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Faf22097401cb951cfd9591d20c59f8270d8f7169))\n* **designer:** config anchor points and outline, jump to section, parse config, highligh ([04b3c52](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F04b3c528a402b391274ef1598416f68dac28dae4))\n* **designer:** config unsaved changes modal ([fd8c588](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ffd8c588cf6f7914a694809b1ad582faac5fbb11f))\n* **designer:** scroll to spot (search incl) ([0c9b707](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0c9b70769a12b8ba4a17ed8f05add8cfe350cfb1))\n* **designer:** search field in config editor ([1a2c2a8](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1a2c2a837daece6085058460765acab64560cb89))\n* **designer:** tab specific section jumping for config editor toggling ([61edd24](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F61edd2445724fa6ccf4567455bf8b1b4857eeabd))\n* **designer:** unsaved changes prompt for config editor ([401fcf7](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F401fcf7091a6e1f857f4dab89feec443609da945))\n* implement delete_project in ProjectService ([1d896b0](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1d896b0c74d600bd9531868dcaeb410722188e68))\n* **server:** add embedding strategies to database api ([#428](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F428)) ([053b965](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F053b9658ace6d4b624c1ca6dab758527396e3e1a))\n* **server:** add mcp_server prop to runtime config ([#415](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F415)) ([693ecad](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F693ecadff546bac8d5eae3a8b79ba03a7ba02f0d))\n* **server:** add project context provider ([#417](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F417)) ([4d50045](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F4d50045841f17fc61c635edcfa43cad565e15e12))\n* **server:** implement delete project api ([db0536b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fdb0536b9035717d932d6827a759eba3fd9deb3c5))\n* update API endpoint to use new delete method ([8e6233d](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F8e6233d2fe2fd7f9e30c30131b5b053cb3686099))\n\n\n### Bug Fixes\n\n* allow deleting projects with invalid configs ([b125192](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fb125192f0512f9e4c6ce44fd489df19fbc36a672))\n* clean up in-memory sessions ([476fbc2](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F476fbc2cee52bab853acad87488328e0627e9942))\n* **cli:** ensure we use a managed python ([#410](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F410)) ([13683ea](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F13683ea810469971e27fea3fb2ee3821f59bc4b8))\n* delete entire directory ([8cfc62d](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F8cfc62d131e606962a93a847395ae6c0f2ff4e0a))\n* **designer:** add footer to chat window to nudge testing chats to test tab ([1ba4fd8](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F1ba4fd8bb21b59603875f6f7bc29aa3cb8b61be0))\n* **designer:** add title to chat window \"build assistant\" ([020080e](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F020080efc106dec44a8a8126013789b801a948d5))\n* **designer:** address comments ([3908d4c](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F3908d4c59889b1208c823a5827d4771d4b0df8fc))\n* **designer:** address comments and slim down PR maybe ([6f465c8](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F6f465c86b75c70f3ca01ce9a87ee867abcbfce09))\n* **designer:** address cubic comment ([6a36728](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F6a3672878d536a612d7ce7a51d7f95c23c9887ea))\n* **designer:** addressing comments ([6f26ae8](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F6f26ae8f89c3f317292379904bdc0c9118abfe09))\n* **designer:** addressing comments ([7f491b6](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F7f491b625cadd08714a1fb0c75e8fae5ccba3be1))\n* **designer:** addressing field validation comment ([fbcf7f5](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ffbcf7f5c1332ae1ab42a8e919622c24387db15f2))\n* **designer:** allow edit database ([b9bddb9](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fb9bddb98f878a0873270d3b305e1397c60da1230))\n* **designer:** allow model swap from existing project and disk models ([#433](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F433)) ([0e5870a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0e5870af54395f9e42c69956498b1089490a2a54))\n* **designer:** allow project delete from designer with new API ([#418](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F418)) ([ae9c6ba](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcomm","2025-11-11T17:13:52",{"id":286,"version":287,"summary_zh":288,"released_at":289},109358,"v0.0.11","## [0.0.11](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.10...v0.0.11) (2025-10-31)\n\n\n### Features\n\n* **designer:** run as part of server ([#389](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F389)) ([72b7942](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F72b7942711b8b8dfdfa7971a817d9295458d38b3))\n\n\n### Bug Fixes\n\n* **designer:** make the chats work ([#387](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F387)) ([f4e3e0b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff4e3e0b33e0adb15acc3ea5a7aac601a1cb99ffb))\n* **runtime:** docker tweak ([54ba31a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F54ba31a611c9334c2c4dbfee5541c878e87f8665))\n\n\n### Miscellaneous Chores\n\n* release 0.0.11 ([f624e6b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff624e6bbb460588f8427b90bc24650f6bb12d54e))","2025-10-31T22:38:48",{"id":291,"version":292,"summary_zh":293,"released_at":294},109359,"v0.0.10","## [0.0.10](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcompare\u002Fv0.0.9...v0.0.10) (2025-10-31)\n\n\n### Features\n\n* add MCP server and improve config importing ([#301](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F301)) ([65fc50f](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F65fc50faeef3a34e62d69d946da36d85e8e8a355))\n* **cli:** add services status command ([#363](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F363)) ([69166ce](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F69166ce3042f00dbda260f70beb44ef4a3698d1a))\n* **cli:** enable upgrades via TUI menu ([#339](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F339)) ([ec80e34](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fec80e34bf887d699be25e7b51514dca1a9816ac9))\n* **designer:** allow models to refernce prompt sets ([0865e3b](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0865e3b6fdc1e8222b7a0d2db60cbd05d856e422))\n* **designer:** double click to toggle to tab home state + polish padding on header ([fa30e2a](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ffa30e2aa19ec5310c11073136b81cc781f3757c5))\n* **designer:** download models and use disk models ([#382](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F382)) ([48c8ee5](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F48c8ee591b7cb480f1f18dc5df73097d07afd934))\n* **designer:** Enhanced dataset processing progress tracking and results display ([#375](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F375)) ([6689606](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F668960633fd2d968d5217be8f93a7734ed76475e))\n* **designer:** hide cloud models, update local models to capture name and description ([7b9cc80](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F7b9cc8008a1dbc42858aedbdba73ba4b6130b11b))\n* **designer:** hook up dev chat to designer chat ([ab56435](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fab56435952faf79b4201ab3dde11059b64d8e8b3))\n* **designer:** hook up models to read and display the correct model ([c594db9](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fc594db93c465d9e2ae6db83992509cb44e9a6b65))\n* **designer:** hook up models to read and display the correct model ([8fb249d](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F8fb249d441dfaa2cc217ef160dc2a616edc4c63f))\n* **designer:** hook up test chat ([#378](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F378)) ([92e4eff](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F92e4eff3b5702868cb4908c04c6838d01002eda2))\n* **designer:** prompts and models ux quick updates so they can be hooked up ([fe47d08](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ffe47d086905024e0c4c3828e009e3e311879d374))\n* **designer:** update \"add prompt\" modal to \"add to set\" ([16d4848](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F16d4848651ebf7953f83a46df371c0aba7e44cf0))\n* **designer:** Updating the Data to be dynamic, update config, and process folders.  ([#361](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F361)) ([bbc69ee](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fbbc69ee57afd474643aa6ffd64279f113348a823))\n* **designer:** UX for prompt sets ([67e3a77](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F67e3a771d79a50b24a682692023f053981c23cdd))\n* editable code in code viewer ([8bf1fe9](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F8bf1fe9a07d8529c8a4a02ba23fe70fd0bf0509f))\n* **rag:** .msg parsing and loader refactor ([#309](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F309)) ([424eadf](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F424eadf6da8d6e86aa2010993bc2eeb10bfa201e))\n* universal runtime and native processes ([#349](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F349)) ([d535842](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fd535842ad0d0ea092de2352711445e46647dc562))\n\n\n### Bug Fixes\n\n* add Python version constraints for PyTorch compatibility ([#380](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F380)) ([df0a9b4](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fdf0a9b466c8ed418cdb6b4e6970ed97d6cd9ae3d))\n* **ci:** bust Go build cache when source files change ([b1f0c24](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Fb1f0c248c4f4d1d032147c8f4de224c7491ba5fa))\n* **ci:** make chmod non-fatal in docker compose test ([2cde502](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F2cde50297526d1b4198f51405196cf77ff9778d4)), closes [#372](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F372)\n* **ci:** resolve docker compose test permission error on self-hosted runners ([f73e4c6](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002Ff73e4c62ed4cf29251fa4184211f7b2dc96ec146)), closes [#367](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F367)\n* **ci:** use sudo to remove docker-created files in test cleanup ([0c3f532](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F0c3f532c2eb34f1cd45c0fc3c5ca3b6d568fe6ea))\n* Clear-text logging of sensitive information ([#345](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fissues\u002F345)) ([5a0f2be](https:\u002F\u002Fgithub.com\u002Fllama-farm\u002Fllamafarm\u002Fcommit\u002F5a0f2bef130a6792bf2099bba982bc498656ad90))\n* **dashboard:** swap out fake metric cards on dashboard with real counts ([f940dd2](https:\u002F\u002Fgithub.com\u002Fllama-f","2025-10-31T04:19:13"]