[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-griptape-ai--griptape":3,"tool-griptape-ai--griptape":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},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,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},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",142651,2,"2026-04-06T23:34:12",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},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。",107888,"2026-04-06T11:32:50",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":10,"last_commit_at":59,"category_tags":60,"status":17},4487,"LLMs-from-scratch","rasbt\u002FLLMs-from-scratch","LLMs-from-scratch 是一个基于 PyTorch 的开源教育项目，旨在引导用户从零开始一步步构建一个类似 ChatGPT 的大型语言模型（LLM）。它不仅是同名技术著作的官方代码库，更提供了一套完整的实践方案，涵盖模型开发、预训练及微调的全过程。\n\n该项目主要解决了大模型领域“黑盒化”的学习痛点。许多开发者虽能调用现成模型，却难以深入理解其内部架构与训练机制。通过亲手编写每一行核心代码，用户能够透彻掌握 Transformer 架构、注意力机制等关键原理，从而真正理解大模型是如何“思考”的。此外，项目还包含了加载大型预训练权重进行微调的代码，帮助用户将理论知识延伸至实际应用。\n\nLLMs-from-scratch 特别适合希望深入底层原理的 AI 开发者、研究人员以及计算机专业的学生。对于不满足于仅使用 API，而是渴望探究模型构建细节的技术人员而言，这是极佳的学习资源。其独特的技术亮点在于“循序渐进”的教学设计：将复杂的系统工程拆解为清晰的步骤，配合详细的图表与示例，让构建一个虽小但功能完备的大模型变得触手可及。无论你是想夯实理论基础，还是为未来研发更大规模的模型做准备",90106,"2026-04-06T11:19:32",[35,15,13,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":76,"owner_email":77,"owner_twitter":78,"owner_website":79,"owner_url":80,"languages":81,"stars":98,"forks":99,"last_commit_at":100,"license":101,"difficulty_score":32,"env_os":102,"env_gpu":102,"env_ram":102,"env_deps":103,"category_tags":106,"github_topics":107,"view_count":32,"oss_zip_url":76,"oss_zip_packed_at":76,"status":17,"created_at":116,"updated_at":117,"faqs":118,"releases":146},4844,"griptape-ai\u002Fgriptape","griptape","Modular Python framework for AI agents and workflows with chain-of-thought reasoning, tools, and memory. ","Griptape 是一款专为简化生成式 AI 应用开发而设计的模块化 Python 框架。它旨在解决开发者在构建复杂 AI 智能体和工作流时面临的挑战，特别是如何有效整合大语言模型（LLM）、检索增强生成（RAG）以及外部工具调用。通过提供清晰灵活的抽象层，Griptape 让开发者能够轻松编排具备“思维链”推理能力的智能系统，并高效管理对话记忆与任务上下文。\n\n这款工具非常适合希望快速原型设计或构建生产级 AI 应用的 Python 开发者及研究人员。其核心亮点在于独特的架构设计：用户可以将任务组装成线性“管道”或并行“工作流”，并利用丰富的“驱动”模块无缝切换不同的模型供应商、向量数据库或多模态服务，而无需修改核心业务逻辑。此外，Griptape 内置了多种实用工具（如网页搜索、文件操作），并支持自定义扩展。对于偏好可视化操作的用户，它还提供了配套的无代码桌面应用 Griptape Nodes。总体而言，Griptape 以严谨的工程化思路，帮助团队更稳健地打造智能化解决方案。","![Griptape](docs\u002Fassets\u002Fimg\u002FGRIPTAPE_from_Foundry_rgb_black.svg)\n\n[![PyPI Version](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fgriptape.svg)](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fgriptape)\n[![Tests](https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Factions\u002Fworkflows\u002Funit-tests.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Factions\u002Fworkflows\u002Funit-tests.yml)\n[![Docs](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fgriptape-ai_griptape_readme_13d664e1afd7.png)](https:\u002F\u002Fgriptape.readthedocs.io\u002F)\n[![Checked with pyright](https:\u002F\u002Fmicrosoft.github.io\u002Fpyright\u002Fimg\u002Fpyright_badge.svg)](https:\u002F\u002Fmicrosoft.github.io\u002Fpyright\u002F)\n[![Ruff](https:\u002F\u002Fimg.shields.io\u002Fendpoint?url=https:\u002F\u002Fraw.githubusercontent.com\u002Fastral-sh\u002Fruff\u002Fmain\u002Fassets\u002Fbadge\u002Fv2.json)](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fruff)\n[![codecov](https:\u002F\u002Fcodecov.io\u002Fgithub\u002Fgriptape-ai\u002Fgriptape\u002Fgraph\u002Fbadge.svg?token=HUBqUpl3NB)](https:\u002F\u002Fcodecov.io\u002Fgithub\u002Fgriptape-ai\u002Fgriptape)\n[![Griptape Discord](https:\u002F\u002Fdcbadge.vercel.app\u002Fapi\u002Fserver\u002FgnWRz88eym?compact=true&style=flat)](https:\u002F\u002Fdiscord.gg\u002Fgriptape)\n\nGriptape is a Python framework designed to simplify the development of generative AI (genAI) applications.\nIt offers a set of straightforward, flexible abstractions for working with areas such as Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and much more.\n\n> **Looking for a no-code experience?** Check out [Griptape Nodes](https:\u002F\u002Fwww.griptapenodes.com\u002Fgriptape-nodes-desktop), a visual desktop application for building and running AI workflows.\n\n## 🛠️ Core Components\n\n### 🏗️ Structures\n\n- 🤖 **Agents** consist of a single Task, configured for Agent-specific behavior.\n- 🔄 **Pipelines** organize a sequence of Tasks so that the output from one Task may flow into the next.\n- 🌐 **Workflows** configure Tasks to operate in parallel.\n\n### 📝 Tasks\n\nTasks are the core building blocks within Structures, enabling interaction with Engines, Tools, and other Griptape components.\n\n### 🧠 Memory\n\n- 💬 **Conversation Memory** enables LLMs to retain and retrieve information across interactions.\n- 🗃️ **Task Memory** keeps large or sensitive Task outputs off the prompt that is sent to the LLM.\n- 📊 **Meta Memory** enables passing in additional metadata to the LLM, enhancing the context and relevance of the interaction.\n\n### 🚗 Drivers\n\nDrivers facilitate interactions with external resources and services in Griptape. \nThey allow you to swap out functionality and providers with minimal changes to your business logic.\n\n#### LLM & Orchestration\n- 🗣️ **Prompt Drivers**: Manage textual and image interactions with LLMs.\n- 🤖 **Assistant Drivers**: Enable interactions with various “assistant” services.\n- 📜 **Ruleset Drivers**: Load and apply rulesets from external sources.\n- 🧠 **Conversation Memory Drivers**: Store and retrieve conversational data.\n- 📡 **Event Listener Drivers**: Forward framework events to external services.\n- 🏗️ **Structure Run Drivers**: Execute structures locally or in the cloud.\n\n#### Retrieval & Storage\n- 🔢 **Embedding Drivers**: Generate vector embeddings from textual inputs.\n- 🔀 **Rerank Drivers**: Rerank search results for improved relevance.\n- 💾 **Vector Store Drivers**: Manage the storage and retrieval of embeddings.\n- 🗂️ **File Manager Drivers**: Handle file operations on local and remote storage.\n- 💼 **SQL Drivers**: Interact with SQL databases.\n\n#### Multimodal\n- 🎨 **Image Generation Drivers**: Create images from text descriptions.\n- 🗣️ **Text to Speech Drivers**: Convert text to speech.\n- 🎙️ **Audio Transcription Drivers**: Convert audio to text.\n\n#### Web\n- 🔍 **Web Search Drivers**: Search the web for information.\n- 🌐 **Web Scraper Drivers**: Extract data from web pages.\n\n#### Observability\n- 📈 **Observability Drivers**: Send trace and event data to observability platforms.\n\n### 🔧 Tools\n\nTools provide capabilities for LLMs to interact with data and services.\nGriptape includes a variety of [built-in Tools](https:\u002F\u002Fdocs.griptape.ai\u002Fstable\u002Fgriptape-framework\u002Ftools\u002Fofficial-tools\u002F), and makes it easy to create [custom Tools](https:\u002F\u002Fdocs.griptape.ai\u002Fstable\u002Fgriptape-framework\u002Ftools\u002Fcustom-tools\u002F).\n\n### 🚂 Engines\n\nEngines wrap Drivers and provide use-case-specific functionality:\n\n- 📊 **RAG Engine** is an abstraction for implementing modular Retrieval Augmented Generation (RAG) pipelines.\n- 🛠️ **Extraction Engine** extracts JSON or CSV data from unstructured text.\n- 📝 **Summary Engine** generates summaries from textual content.\n- ✅ **Eval Engine** evaluates and scores the quality of generated text.\n\n### 📦 Additional Components\n\n- 📐 **Rulesets** steer LLM behavior with minimal prompt engineering.\n- 🔄 **Loaders** load data from various sources.\n- 🏺 **Artifacts** allow for passing data of different types between Griptape components.\n- ✂️ **Chunkers** segment texts into manageable pieces for diverse text types.\n- 🔢 **Tokenizers** count the number of tokens in a text to not exceed LLM token limits.\n\n## Documentation\n\nPlease visit the [docs](https:\u002F\u002Fdocs.griptape.ai\u002F) for information on installation and usage.\n\nCheck out [Griptape Trade School](https:\u002F\u002Flearn.griptape.ai\u002F) for free online courses.\n\n## Hello World Example\n\nHere's a minimal example of griptape:\n\n\n```python\nfrom griptape.drivers.prompt.openai import OpenAiChatPromptDriver\nfrom griptape.rules import Rule\nfrom griptape.tasks import PromptTask\n\ntask = PromptTask(\n    prompt_driver=OpenAiChatPromptDriver(model=\"gpt-4.1\"),\n    rules=[Rule(\"Keep your answer to a few sentences.\")],\n)\n\nresult = task.run(\"How do I do a kickflip?\")\n\nprint(result.value)\n```\n\n```text\nTo do a kickflip, start by positioning your front foot slightly angled near the middle of the board and your back foot on the tail.\nPop the tail down with your back foot while flicking the edge of the board with your front foot to make it spin.\nJump and keep your body centered over the board, then catch it with your feet and land smoothly. Practice and patience are key!\n```\n\n\n## Task and Workflow Example\n\nHere is a concise example using griptape to research open source projects:\n\n```python\nfrom griptape.drivers.prompt.openai_chat_prompt_driver import OpenAiChatPromptDriver\nfrom griptape.drivers.web_search.duck_duck_go import DuckDuckGoWebSearchDriver\nfrom griptape.rules import Rule, Ruleset\nfrom griptape.structures import Workflow\nfrom griptape.tasks import PromptTask, TextSummaryTask\nfrom griptape.tools import WebScraperTool, WebSearchTool\nfrom griptape.utils import StructureVisualizer\nfrom pydantic import BaseModel\n\n\nclass Feature(BaseModel):\n    name: str\n    description: str\n    emoji: str\n\n\nclass Output(BaseModel):\n    answer: str\n    key_features: list[Feature]\n\n\nprojects = [\"griptape\", \"langchain\", \"crew-ai\", \"pydantic-ai\"]\n\nprompt_driver = OpenAiChatPromptDriver(model=\"gpt-4.1\")\nworkflow = Workflow(\n    tasks=[\n        [\n            PromptTask(\n                id=f\"project-{project}\",\n                input=\"Tell me about the open source project: {{ project }}.\",\n                prompt_driver=prompt_driver,\n                context={\"project\": projects},\n                output_schema=Output,\n                tools=[\n                    WebSearchTool(\n                        web_search_driver=DuckDuckGoWebSearchDriver(),\n                    ),\n                    WebScraperTool(),\n                ],\n                child_ids=[\"summary\"],\n            )\n            for project in projects\n        ],\n        TextSummaryTask(\n            input=\"{{ parents_output_text }}\",\n            id=\"summary\",\n            rulesets=[\n                Ruleset(\n                    name=\"Format\", rules=[Rule(\"Be detailed.\"), Rule(\"Include emojis.\")]\n                )\n            ],\n        ),\n    ]\n)\n\nworkflow.run()\n\nprint(StructureVisualizer(workflow).to_url())\n```\n\n```text\n Output: Here's a detailed summary of the open-source projects mentioned:\n\n 1. **Griptape** 🛠️:                                                                                                            \n    - Griptape is a modular Python framework designed for creating AI-powered applications. It focuses on securely connecting to\n enterprise data and APIs. The framework provides structured components like Agents, Pipelines, and Workflows, allowing for both\n parallel and sequential operations. It includes built-in tools and supports custom tool creation for data and service\n interaction.\n\n 2. **LangChain** 🔗:\n    - LangChain is a framework for building applications powered by Large Language Models (LLMs). It offers a standard interface\n for models, embeddings, and vector stores, facilitating real-time data augmentation and model interoperability. LangChain\n integrates with various data sources and external systems, making it adaptable to evolving technologies.\n\n 3. **CrewAI** 🤖:\n    - CrewAI is a standalone Python framework for orchestrating multi-agent AI systems. It allows developers to create and\n manage AI agents that collaborate on complex tasks. CrewAI emphasizes ease of use and scalability, providing tools and\n documentation to help developers build AI-powered solutions.\n\n 4. **Pydantic-AI** 🧩:\n    - Pydantic-AI is a Python agent framework that simplifies the development of production-grade applications with Generative\n AI. Built on Pydantic, it supports various AI models and provides features like type-safe design, structured response\n validation, and dependency injection. Pydantic-AI aims to bring the ease of FastAPI development to AI applications.\n\n These projects offer diverse tools and frameworks for developing AI applications, each with unique features and capabilities\n tailored to different aspects of AI development.\n```\n\n```mermaid\n    graph TD;\n    griptape-->summary;\n    langchain-->summary;\n    pydantic-ai-->summary;\n    crew-ai-->summary;\n```\n\n## Versioning\n\nGriptape uses [Semantic Versioning](https:\u002F\u002Fsemver.org\u002F).\n\n## Contributing\n\nThank you for considering contributing to Griptape! Before you start, please review our [Contributing Guidelines](https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fblob\u002Fmain\u002FCONTRIBUTING.md).\n\n## License\n\nGriptape is available under the Apache 2.0 License.\n","![Griptape](docs\u002Fassets\u002Fimg\u002FGRIPTAPE_from_Foundry_rgb_black.svg)\n\n[![PyPI Version](https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fgriptape.svg)](https:\u002F\u002Fpypi.python.org\u002Fpypi\u002Fgriptape)\n[![Tests](https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Factions\u002Fworkflows\u002Funit-tests.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Factions\u002Fworkflows\u002Funit-tests.yml)\n[![Docs](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fgriptape-ai_griptape_readme_13d664e1afd7.png)](https:\u002F\u002Fgriptape.readthedocs.io\u002F)\n[![Checked with pyright](https:\u002F\u002Fmicrosoft.github.io\u002Fpyright\u002Fimg\u002Fpyright_badge.svg)](https:\u002F\u002Fmicrosoft.github.io\u002Fpyright\u002F)\n[![Ruff](https:\u002F\u002Fimg.shields.io\u002Fendpoint?url=https:\u002F\u002Fraw.githubusercontent.com\u002Fastral-sh\u002Fruff\u002Fmain\u002Fassets\u002Fbadge\u002Fv2.json)](https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fruff)\n[![codecov](https:\u002F\u002Fcodecov.io\u002Fgithub\u002Fgriptape-ai\u002Fgriptape\u002Fgraph\u002Fbadge.svg?token=HUBqUpl3NB)](https:\u002F\u002Fcodecov.io\u002Fgithub\u002Fgriptape-ai\u002Fgriptape)\n[![Griptape Discord](https:\u002F\u002Fdcbadge.vercel.app\u002Fapi\u002Fserver\u002FgnWRz88eym?compact=true&style=flat)](https:\u002F\u002Fdiscord.gg\u002Fgriptape)\n\nGriptape 是一个 Python 框架，旨在简化生成式 AI（genAI）应用程序的开发。它提供了一组简单灵活的抽象层，用于处理大型语言模型（LLM）、检索增强生成（RAG）等领域。\n\n> **寻找无代码体验？** 请查看 [Griptape Nodes](https:\u002F\u002Fwww.griptapenodes.com\u002Fgriptape-nodes-desktop)，这是一款用于构建和运行 AI 工作流的可视化桌面应用。\n\n## 🛠️ 核心组件\n\n### 🏗️ 结构\n\n- 🤖 **代理** 由单个任务组成，配置为特定于代理的行为。\n- 🔄 **管道** 将一系列任务组织起来，使前一个任务的输出可以作为下一个任务的输入。\n- 🌐 **工作流** 配置任务以并行运行。\n\n### 📝 任务\n\n任务是结构中的核心构建块，能够与引擎、工具及其他 Griptape 组件交互。\n\n### 🧠 内存\n\n- 💬 **对话内存** 使 LLM 能够在多次交互中保持和检索信息。\n- 🗃️ **任务内存** 可以将较大或敏感的任务输出从发送给 LLM 的提示中移除。\n- 📊 **元内存** 允许向 LLM 传递额外的元数据，从而增强交互的上下文和相关性。\n\n### 🚗 驱动程序\n\n驱动程序促进了 Griptape 中与外部资源和服务的交互。它们允许您在对业务逻辑进行最少更改的情况下替换功能和提供商。\n\n#### LLM 与编排\n- 🗣️ **提示驱动程序**：管理与 LLM 的文本和图像交互。\n- 🤖 **助手驱动程序**：支持与各种“助手”服务的交互。\n- 📜 **规则集驱动程序**：从外部来源加载并应用规则集。\n- 🧠 **对话内存驱动程序**：存储和检索对话数据。\n- 📡 **事件监听器驱动程序**：将框架事件转发到外部服务。\n- 🏗️ **结构运行驱动程序**：在本地或云端执行结构。\n\n#### 检索与存储\n- 🔢 **嵌入驱动程序**：根据文本输入生成向量嵌入。\n- 🔀 **重排序驱动程序**：对搜索结果进行重新排序以提高相关性。\n- 💾 **向量存储驱动程序**：管理嵌入的存储和检索。\n- 🗂️ **文件管理驱动程序**：处理本地和远程存储上的文件操作。\n- 💼 **SQL 驱动程序**：与 SQL 数据库交互。\n\n#### 多模态\n- 🎨 **图像生成驱动程序**：根据文本描述创建图像。\n- 🗣️ **文本转语音驱动程序**：将文本转换为语音。\n- 🎙️ **音频转录驱动程序**：将音频转换为文本。\n\n#### 网络\n- 🔍 **网页搜索驱动程序**：在网络上搜索信息。\n- 🌐 **网页抓取驱动程序**：从网页中提取数据。\n\n#### 可观测性\n- 📈 **可观测性驱动程序**：将跟踪和事件数据发送到可观测性平台。\n\n### 🔧 工具\n\n工具为 LLM 提供了与数据和服务交互的能力。Griptape 包含多种[内置工具](https:\u002F\u002Fdocs.griptape.ai\u002Fstable\u002Fgriptape-framework\u002Ftools\u002Fofficial-tools\u002F)，并且也便于创建[自定义工具](https:\u002F\u002Fdocs.griptape.ai\u002Fstable\u002Fgriptape-framework\u002Ftools\u002Fcustom-tools\u002F)。\n\n### 🚂 引擎\n\n引擎封装了驱动程序，并提供了针对特定用例的功能：\n\n- 📊 **RAG 引擎** 是实现模块化检索增强生成（RAG）管道的抽象层。\n- 🛠️ **提取引擎** 从非结构化文本中提取 JSON 或 CSV 数据。\n- 📝 **摘要引擎** 生成文本内容的摘要。\n- ✅ **评估引擎** 对生成文本的质量进行评估和打分。\n\n### 📦 其他组件\n\n- 📐 **规则集** 在尽量减少提示工程的情况下引导 LLM 的行为。\n- 🔄 **加载器** 从各种来源加载数据。\n- 🏺 **工件** 允许在 Griptape 组件之间传递不同类型的数据。\n- ✂️ **分块器** 将不同类型的文本分割成易于处理的小段。\n- 🔢 **分词器** 计算文本中的标记数量，以确保不超过 LLM 的标记限制。\n\n## 文档\n\n有关安装和使用的信息，请访问 [文档](https:\u002F\u002Fdocs.griptape.ai\u002F)。\n\n欢迎访问 [Griptape 技术学校](https:\u002F\u002Flearn.griptape.ai\u002F)，获取免费在线课程。\n\n## Hello World 示例\n\n以下是 Griptape 的一个最小示例：\n\n\n```python\nfrom griptape.drivers.prompt.openai import OpenAiChatPromptDriver\nfrom griptape.rules import Rule\nfrom griptape.tasks import PromptTask\n\ntask = PromptTask(\n    prompt_driver=OpenAiChatPromptDriver(model=\"gpt-4.1\"),\n    rules=[Rule(\"Keep your answer to a few sentences.\")],\n)\n\nresult = task.run(\"How do I do a kickflip?\")\n\nprint(result.value)\n```\n\n```text\n要完成一个踢翻动作，首先将前脚稍微倾斜地放在板子中间附近，后脚踩在板尾上。\n用后脚向下弹起板尾，同时用前脚轻踢板边使其旋转。\n跳起来并保持身体位于板子中央，然后用双脚接住滑板并平稳落地。练习和耐心是关键！\n```\n\n## 任务与工作流示例\n\n以下是一个使用 Griptape 研究开源项目的简洁示例：\n\n```python\nfrom griptape.drivers.prompt.openai_chat_prompt_driver import OpenAiChatPromptDriver\nfrom griptape.drivers.web_search.duck_duck_go import DuckDuckGoWebSearchDriver\nfrom griptape.rules import Rule, Ruleset\nfrom griptape.structures import Workflow\nfrom griptape.tasks import PromptTask, TextSummaryTask\nfrom griptape.tools import WebScraperTool, WebSearchTool\nfrom griptape.utils import StructureVisualizer\nfrom pydantic import BaseModel\n\n\nclass Feature(BaseModel):\n    name: str\n    description: str\n    emoji: str\n\n\nclass Output(BaseModel):\n    answer: str\n    key_features: list[Feature]\n\n\nprojects = [\"griptape\", \"langchain\", \"crew-ai\", \"pydantic-ai\"]\n\nprompt_driver = OpenAiChatPromptDriver(model=\"gpt-4.1\")\nworkflow = Workflow(\n    tasks=[\n        [\n            PromptTask(\n                id=f\"project-{project}\",\n                input=\"Tell me about the open source project: {{ project }}.\",\n                prompt_driver=prompt_driver,\n                context={\"project\": projects},\n                output_schema=Output,\n                tools=[\n                    WebSearchTool(\n                        web_search_driver=DuckDuckGoWebSearchDriver(),\n                    ),\n                    WebScraperTool(),\n                ],\n                child_ids=[\"summary\"],\n            )\n            for project in projects\n        ],\n        TextSummaryTask(\n            input=\"{{ parents_output_text }}\",\n            id=\"summary\",\n            rulesets=[\n                Ruleset(\n                    name=\"Format\", rules=[Rule(\"Be detailed.\"), Rule(\"Include emojis.\")]\n                )\n            ],\n        ),\n    ]\n)\n\nworkflow.run()\n\nprint(StructureVisualizer(workflow).to_url())\n```\n\n```text\n输出：以下是所提及的开源项目的详细总结：\n\n1. **Griptape** 🛠️：\n   - Griptape 是一个模块化的 Python 框架，专为构建 AI 驱动的应用程序而设计。它专注于安全地连接企业数据和 API。该框架提供了 Agent、Pipeline 和 Workflow 等结构化组件，支持并行和串行操作。Griptape 内置了多种工具，并支持自定义工具的创建，以实现与数据和服务的交互。\n\n2. **LangChain** 🔗：\n   - LangChain 是一个用于构建由大型语言模型（LLM）驱动的应用程序的框架。它为模型、嵌入和向量存储提供了标准接口，便于实时数据增强和模型之间的互操作性。LangChain 可以集成多种数据源和外部系统，使其能够适应不断发展的技术。\n\n3. **CrewAI** 🤖：\n   - CrewAI 是一个独立的 Python 框架，用于编排多智能体 AI 系统。它允许开发者创建和管理协作完成复杂任务的 AI 智能体。CrewAI 强调易用性和可扩展性，提供工具和文档来帮助开发者构建 AI 驱动的解决方案。\n\n4. **Pydantic-AI** 🧩：\n   - Pydantic-AI 是一个基于 Pydantic 的 Python 智能体框架，旨在简化生成式 AI 生产级应用程序的开发。它支持多种 AI 模型，并提供类型安全设计、结构化响应验证和依赖注入等功能。Pydantic-AI 试图将 FastAPI 开发的便捷性引入到 AI 应用中。\n\n这些项目提供了多样化的工具和框架来开发 AI 应用程序，每个项目都有其独特的功能和能力，分别针对 AI 开发的不同方面。\n```\n\n```mermaid\n    graph TD;\n    griptape-->summary;\n    langchain-->summary;\n    pydantic-ai-->summary;\n    crew-ai-->summary;\n```\n\n## 版本控制\n\nGriptape 使用 [语义化版本控制](https:\u002F\u002Fsemver.org\u002F)。\n\n## 贡献\n\n感谢您考虑为 Griptape 做出贡献！在开始之前，请先阅读我们的 [贡献指南](https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fblob\u002Fmain\u002FCONTRIBUTING.md)。\n\n## 许可证\n\nGriptape 采用 Apache 2.0 许可证授权。","# Griptape 快速上手指南\n\nGriptape 是一个专为简化生成式 AI（GenAI）应用开发而设计的 Python 框架。它提供了一套灵活且直观的抽象组件，适用于大语言模型（LLM）、检索增强生成（RAG）等场景。\n\n## 环境准备\n\n在开始之前，请确保您的开发环境满足以下要求：\n\n- **操作系统**：Linux、macOS 或 Windows\n- **Python 版本**：Python 3.9 或更高版本\n- **前置依赖**：\n  - `pip` (Python 包管理工具)\n  - 有效的 LLM API 密钥（如 OpenAI API Key），用于驱动模型运行\n\n> **提示**：国内开发者若遇到网络问题，建议在安装时配置国内镜像源（如清华源或阿里源）以加速下载。\n\n## 安装步骤\n\n使用 `pip` 安装 Griptape 核心库：\n\n```bash\npip install griptape\n```\n\n**推荐使用国内镜像源加速安装：**\n\n```bash\npip install griptape -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n```\n\n如需使用特定功能（如网页搜索、向量数据库等），可安装额外依赖：\n\n```bash\npip install \"griptape[all]\"\n```\n\n## 基本使用\n\n以下是一个最简单的 \"Hello World\" 示例，展示如何使用 Griptape 创建一个简单的提示任务并与 LLM 交互。\n\n### 1. 编写代码\n\n创建文件 `hello_griptape.py`，并填入以下代码：\n\n```python\nfrom griptape.drivers.prompt.openai import OpenAiChatPromptDriver\nfrom griptape.rules import Rule\nfrom griptape.tasks import PromptTask\n\n# 初始化提示任务\ntask = PromptTask(\n    prompt_driver=OpenAiChatPromptDriver(model=\"gpt-4.1\"),\n    rules=[Rule(\"Keep your answer to a few sentences.\")],\n)\n\n# 运行任务\nresult = task.run(\"How do I do a kickflip?\")\n\n# 输出结果\nprint(result.value)\n```\n\n> **注意**：运行前请确保已在环境变量中配置好 `OPENAI_API_KEY`，或者在代码中显式传入 API Key。\n\n### 2. 运行程序\n\n在终端执行脚本：\n\n```bash\npython hello_griptape.py\n```\n\n### 3. 预期输出\n\n程序将调用 LLM 并返回类似以下的回答：\n\n```text\nTo do a kickflip, start by positioning your front foot slightly angled near the middle of the board and your back foot on the tail.\nPop the tail down with your back foot while flicking the edge of the board with your front foot to make it spin.\nJump and keep your body centered over the board, then catch it with your feet and land smoothly. Practice and patience are key!\n```\n\n---\n\n通过以上步骤，您已成功运行了第一个 Griptape 应用。接下来您可以探索 **Structures**（如 Agents, Pipelines, Workflows）和 **Tools** 来构建更复杂的 AI 工作流。","某电商公司的数据团队需要构建一个能自动分析用户评论、检索历史客诉记录并生成个性化回复的智能客服助手。\n\n### 没有 griptape 时\n- 开发者需手动编写大量胶水代码来串联 LLM 调用、向量数据库检索和网络搜索工具，导致逻辑耦合严重且难以维护。\n- 缺乏统一的记忆管理机制，AI 无法在长对话中准确引用用户之前的投诉细节，常出现“失忆”或重复询问的情况。\n- 当需要切换底层大模型供应商（如从 OpenAI 切换到 Anthropic）或更换向量存储方案时，必须重构核心业务逻辑，迁移成本极高。\n- 复杂的任务流程（如先搜索再总结最后生成回复）难以编排，并行处理多个用户请求时系统稳定性差，调试链路追踪困难。\n\n### 使用 griptape 后\n- 利用 griptape 的 Pipeline 和 Workflow 结构，通过声明式代码轻松编排“检索 - 推理 - 生成”的复杂链条，大幅减少样板代码。\n- 内置的 Conversation Memory 和 Task Memory 组件自动管理上下文，确保 AI 能精准记住用户历史交互细节，同时避免敏感数据泄露给模型。\n- 借助 Driver 抽象层，仅需修改配置文件即可无缝切换 LLM 提供商或向量数据库，核心业务逻辑无需任何改动。\n- 原生支持可观测性驱动，自动将任务执行轨迹和事件发送至监控平台，让并行任务的性能瓶颈和错误定位一目了然。\n\ngriptape 通过模块化的架构设计，将原本繁琐的 AI 应用开发转化为灵活的积木搭建过程，显著提升了智能代理的构建效率与系统鲁棒性。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fgriptape-ai_griptape_0f62f72c.png","griptape-ai","Griptape","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fgriptape-ai_8077de91.png","",null,"hello@griptape.ai","griptapeai","https:\u002F\u002Fwww.griptape.ai\u002F","https:\u002F\u002Fgithub.com\u002Fgriptape-ai",[82,86,90,94],{"name":83,"color":84,"percentage":85},"Python","#3572A5",99.4,{"name":87,"color":88,"percentage":89},"Jinja","#a52a22",0.4,{"name":91,"color":92,"percentage":93},"Makefile","#427819",0.1,{"name":95,"color":96,"percentage":97},"Dockerfile","#384d54",0,2508,228,"2026-04-06T23:36:44","Apache-2.0","未说明",{"notes":104,"python":102,"dependencies":105},"README 中未明确列出具体的操作系统、GPU、内存或 Python 版本要求。该工具是一个用于构建生成式 AI 应用的 Python 框架，核心功能依赖于外部驱动（如 OpenAI、DuckDuckGo 等）而非本地大型模型推理，因此通常不需要特定的 GPU 配置。具体依赖项需根据实际使用的驱动和功能动态安装。建议查阅官方文档 (docs.griptape.ai) 获取详细的安装指南。",[64],[13,14,15,35],[108,109,110,111,112,113,114,115],"gpt","llm","openai","ai","python","huggingface","anthropic","claude","2026-03-27T02:49:30.150509","2026-04-07T13:28:53.653731",[119,124,129,134,138,142],{"id":120,"question_zh":121,"answer_zh":122,"source_url":123},22035,"为什么 Agent 在使用 gpt-3.5-turbo 时性能下降或无法正常工作？","Griptape 默认更新为使用 gpt-4 以确保 Agent 能 100% 正常运行。如果必须使用 gpt-3.5-turbo 且遇到工具记忆（tool memory）相关的问题，可以通过在初始化 Agent 时禁用工具记忆作为临时解决方案：\n\n```python\nAgent(\n    tool_memory=None\n)\n```\n\n维护者指出 gpt-3.5 的相关问题仍在修复中，建议优先使用 gpt-4。","https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fissues\u002F114",{"id":125,"question_zh":126,"answer_zh":127,"source_url":128},22036,"WebLoader 或 WebScraper 在提取网页信息时为什么会提前停止或遗漏内容？","这通常是由提示词（prompt）的微小变化引起的。维护者确认最近的代码提交已修复了该问题。如果遇到此问题，请确保升级到最新版本（如 v0.20.0 或更高），或者检查是否使用了过时的版本（如 v0.15.3 之前）。此外，对于列表类数据的提取，尝试使用 CSV 提取引擎可能比默认的文本提取效果更好。","https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fissues\u002F146",{"id":130,"question_zh":131,"answer_zh":132,"source_url":133},22037,"为什么 FileManagerTool 无法让 LLM 描述或处理图像文件？","这是因为当前的 LLM 提供商（如 OpenAI）限制了在工具调用（tool calls）中直接传递图像数据的能力。虽然 FileManagerTool 配置了 ImageLoader，但当 Agent 尝试将加载的图像作为工具输出返回给 LLM 时，会收到类似\"Image URLs are only allowed for messages with role 'user'\"的错误。目前这是一个平台限制，需等待 OpenAI 等提供商允许在工具响应中包含图像 URL 或数据。","https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fissues\u002F1432",{"id":135,"question_zh":136,"answer_zh":137,"source_url":128},22038,"如何监控和统计 Griptape Agent 运行时的 Token 消耗？","可以通过注册事件监听器来监控 StartPromptEvent 和 FinishPromptEvent。以下是一个示例代码，用于统计 Token 数量：\n\n```python\nfrom griptape.events import StartPromptEvent, FinishPromptEvent\nfrom griptape import utils\n\ndef handle_prompt_event(event, token_counter):\n    print(event)\n    token_counter.add_tokens(event.token_count)\n\ntoken_counter = utils.TokenCounter()\n\npipeline = Pipeline(\n    # ... 其他配置 ...\n    event_listeners={\n        StartPromptEvent: [\n            lambda e: handle_prompt_event(e, token_counter)\n        ],\n        FinishPromptEvent: [\n            lambda e: handle_prompt_event(e, token_counter)\n        ]\n    }\n)\n```",{"id":139,"question_zh":140,"answer_zh":141,"source_url":128},22039,"如何配置 Agent 以使用特定的模型（如 gpt-4 或 gemini）？","在初始化 Agent 或 Pipeline 时，通过 prompt_driver 参数指定模型。例如，使用 OpenAI 的 gpt-4：\n\n```python\nfrom griptape.drivers import OpenAiChatPromptDriver\nfrom griptape.structures import Agent\n\nagent = Agent(\n    prompt_driver=OpenAiChatPromptDriver(\n        model=\"gpt-4\",\n    ),\n    tools=[...]\n)\n```\n\n如果使用 Google Gemini：\n\n```python\nfrom griptape.drivers import GooglePromptDriver\n\nagent = Agent(\n    prompt_driver=GooglePromptDriver(\n        api_key=GOOGLE_API_KEY, \n        model=\"gemini-2.0-flash-exp\"\n    ),\n    tools=[...]\n)\n```",{"id":143,"question_zh":144,"answer_zh":145,"source_url":123},22040,"如何使用 KnowledgeBaseClient 构建基于本地文档的问答系统？","首先使用 Loader（如 TextLoader, PdfLoader）加载文档，然后将其存入向量存储引擎，最后初始化 KnowledgeBaseClient。示例代码如下：\n\n```python\nfrom griptape.loaders import TextLoader\nfrom griptape.engines import VectorQueryEngine\nfrom griptape.tools import KnowledgeBaseClient\n\nengine = VectorQueryEngine()\nnamespace = \"my-knowledge-base\"\n\n# 加载文本并写入向量库\nartifacts = TextLoader().load(\"path\u002Fto\u002Fyour\u002Fdocument.txt\")\nengine.vector_store_driver.upsert_text_artifacts({namespace: artifacts})\n\n# 创建客户端工具\nkb_client = KnowledgeBaseClient(\n    description=\"包含特定领域知识的知识库。\",\n    query_engine=engine,\n    namespace=namespace\n)\n\n# 将 kb_client 添加到 Agent 的工具列表中\n```\n\n注意：如果在旧版本中遇到性能问题，请尝试升级 Griptape 或暂时禁用 tool_memory。",[147,152,157,162,167,172,177,182,187,192,197,202,207,212,217,222,227,232,237,242],{"id":148,"version":149,"summary_zh":150,"released_at":151},135698,"v1.9.4","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置在 v1.9.4 版本生成 -->\n\n\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.9.3...v1.9.4","2026-03-13T15:38:58",{"id":153,"version":154,"summary_zh":155,"released_at":156},135699,"v1.9.3","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置生成，版本为 v1.9.3 -->\n\n\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.9.2...v1.9.3","2026-02-19T20:24:23",{"id":158,"version":159,"summary_zh":160,"released_at":161},135700,"v1.9.2","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置在 v1.9.2 版本生成 -->\n\n\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.9.1...v1.9.2","2026-02-18T17:52:10",{"id":163,"version":164,"summary_zh":165,"released_at":166},135701,"v1.9.1","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置在 v1.9.1 版本生成 -->\r\n\r\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.9.0...v1.9.1","2026-02-15T00:08:53",{"id":168,"version":169,"summary_zh":170,"released_at":171},135702,"v1.9.0","\u003C!-- 使用 .github\u002Frelease.yml 中的配置在 v1.9.0 版本生成的发布说明 -->\n\n## 变更内容\n### ✨ 功能增强\n* 新增 (drivers-web-scraper-proxy): 向 `ProxyWebScraperDriver` 添加 `_extract_page` 参数，用于在运行时覆盖网页提取逻辑。由 @vachillo 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1935 中实现。\n* 为 agent 添加 `max_subtasks` 参数，由 @ian-griptape-ai 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1973 中实现。\n### 🐛 问题修复\n* ci: 移除 release-please，改用 GitHub 发布功能，由 @collindutter 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1933 中完成。\n* 修复：正确序列化 `GrokPromptDriver` 和 `CoherePromptDriver`，由 @collindutter 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1934 中实现。\n* 在 `BaseVectorStoreDriver` 中始终执行向量的 upsert 操作，由 @cjkindel 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1940 中完成。\n* 修复：为 `OpenAiImageGenerationDriver` 添加缺失的 `ignored_exception_types` 配置，由 @vachillo 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1947 中实现。\n* 修复：检测 mcp 工具是否已处于事件循环中，由 @collindutter 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1994 中完成。\n### 📦 依赖更新\n* 依赖：将生产依赖组中的 8 个依赖项升级，由 @dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1939 中完成。\n* 依赖：将生产依赖组中的 10 个依赖项升级，由 @dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1945 中完成。\n* 构建（依赖）：将 actions\u002Fcheckout 从版本 4 升级到 5，由 @dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1989 中完成。\n* 构建（依赖）：将 tj-actions\u002Fchanged-files 从版本 46 升级到 47，由 @dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F2000 中完成。\n* 构建（依赖）：将 peter-evans\u002Fcreate-pull-request 从版本 7 升级到 8，由 @dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F2044 中完成。\n* 构建（依赖）：将 actions\u002Fcheckout 从版本 5 升级到 6，由 @dependabot[bot] 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F2040 中完成。\n### 其他变更\n* 修复：更新 `OpenAiImageGenerationDriver` 中 gpt-image-1 的文档字符串，由 @ian-griptape-ai 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1936 中完成。\n* ci: 清理稳定版发布流水线，由 @collindutter 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1937 中完成。\n* 添加 `AudioUrlArtifact`，由 @zachgiordano 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1938 中实现。\n* 文档化如何在 Griptape Cloud 中创建基于 ZIP 的工具和结构，由 @cjkindel 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1951 中完成。\n* 修复创建结构\u002F工具时的头部信息，由 @cjkindel 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1953 中完成。\n* 修复 (drivers-prompt-google): 正确处理调试日志，由 @collindutter 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1952 中完成。\n* 支持 Griptape Cloud 的 base_url 带路径访问，由 @cjkindel 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1954 中完成。\n* 杂项：在 main 分支上将版本号递增至下一个次版本，由 @vachillo 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1958 中完成。\n* 新增功能：支持 GriptapeCloudImageGenerationDriver 中的 gpt-image-1，由 @cjkindel 在 https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1964 中实现。\n* 哈","2026-01-13T21:22:41",{"id":173,"version":174,"summary_zh":175,"released_at":176},135703,"v1.8.13","\u003C!-- 发行说明由 .github\u002Frelease.yml 中的配置在 v1.8.13 版本生成 -->\n\n\n\n**完整变更日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.12...v1.8.13","2025-11-24T21:39:07",{"id":178,"version":179,"summary_zh":180,"released_at":181},135704,"v1.8.12","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置生成，版本为 v1.8.12 -->\n\n\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.11...v1.8.12","2025-10-29T23:43:23",{"id":183,"version":184,"summary_zh":185,"released_at":186},135705,"v1.8.11","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置在 v1.8.11 版本生成 -->\n\n\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.10...v1.8.11","2025-10-29T17:04:10",{"id":188,"version":189,"summary_zh":190,"released_at":191},135706,"v1.8.10","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置生成，版本为 v1.8.10 -->\n\n\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.9...v1.8.10","2025-10-28T23:06:47",{"id":193,"version":194,"summary_zh":195,"released_at":196},135707,"v1.8.9","\u003C!-- 发布说明由 .github\u002Frelease.yml 中的配置生成，版本为 v1.8.9 -->\n\n\n\n**完整更新日志**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.8...v1.8.9","2025-10-28T19:45:08",{"id":198,"version":199,"summary_zh":200,"released_at":201},135708,"v1.8.8","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.8 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.7...v1.8.8","2025-10-10T00:35:08",{"id":203,"version":204,"summary_zh":205,"released_at":206},135709,"v1.8.7","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.7 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.6...v1.8.7","2025-10-09T16:18:40",{"id":208,"version":209,"summary_zh":210,"released_at":211},135710,"v1.8.6","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.6 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.5...v1.8.6","2025-10-08T19:01:21",{"id":213,"version":214,"summary_zh":215,"released_at":216},135711,"v1.8.5","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.5 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.4...v1.8.5","2025-10-06T17:15:04",{"id":218,"version":219,"summary_zh":220,"released_at":221},135712,"v1.8.4","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.4 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.3...v1.8.4","2025-09-30T16:35:32",{"id":223,"version":224,"summary_zh":225,"released_at":226},135713,"v1.8.3","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.3 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.2...v1.8.3","2025-09-29T16:32:24",{"id":228,"version":229,"summary_zh":230,"released_at":231},135714,"v1.8.2","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.2 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.8.1...v1.8.2","2025-08-27T21:57:48",{"id":233,"version":234,"summary_zh":235,"released_at":236},135715,"v1.8.1","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.1 -->\n\n## What's Changed\n### ✨ Enhancements\n* feat(drivers-web-scraper-proxy): add `_extract_page` parameter to `ProxyWebScraperDriver` to override web page extraction at runtime. by @vachillo in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1935\n* add max_subtasks to agent by @ian-griptape-ai in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1973\n### 🐛 Bug Fixes\n* ci: remove release-please in favor of github releasing by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1933\n* fix: properly serialize GrokPromptDriver and CoherePromptDriver by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1934\n* Always upsert vectors in BaseVectorStoreDriver by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1940\n* fix: add missing `ignored_exception_types` to `OpenAiImageGenerationDriver` by @vachillo in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1947\n### 📦 Dependencies\n* deps: bump the production-dependencies group with 8 updates by @dependabot[bot] in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1939\n* deps: bump the production-dependencies group with 10 updates by @dependabot[bot] in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1945\n### Other Changes\n* fix: doc string update for gpt-image-1 in OpenAiImageGenerationDriver by @ian-griptape-ai in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1936\n* ci: clean up stable release pipeline by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1937\n* Add AudioUrlArtifact by @zachgiordano in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1938\n* Document how to create ZIP-based tools and structures in Griptape Cloud by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1951\n* Fix headers for create structure\u002Ftool by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1953\n* fix(drivers-prompt-google): correctly handle debug logs by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1952\n* Support Griptape Cloud base_url with paths by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1954\n* chore: bump to next minor version in main branch by @vachillo in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1958\n* feat: Support gpt-image-1 in GriptapeCloudImageGenerationDriver by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1964\n* Handle embedded text with commas for PGAI Vector Store Driver by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1967\n* feat: MCP Tool by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1962\n* Remove dependency on active event_loop in MCP tool initialization by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1969\n* Docs\u002Fmcp by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1972\n* Super __attrs_post_init__ by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1974\n* chore: bump v1.8.0 by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1976\n* feat: Handle event stream error messages in GriptapeCloudPromptDriver by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1983\n* fix: add support for gpt-5 by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1985\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.7.0...v1.8.1","2025-08-08T22:11:06",{"id":238,"version":239,"summary_zh":240,"released_at":241},135716,"v1.8.0","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.8.0 -->\n\n## What's Changed\n### ✨ Enhancements\n* feat(drivers-web-scraper-proxy): add `_extract_page` parameter to `ProxyWebScraperDriver` to override web page extraction at runtime. by @vachillo in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1935\n* add max_subtasks to agent by @ian-griptape-ai in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1973\n### 🐛 Bug Fixes\n* ci: remove release-please in favor of github releasing by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1933\n* fix: properly serialize GrokPromptDriver and CoherePromptDriver by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1934\n* Always upsert vectors in BaseVectorStoreDriver by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1940\n* fix: add missing `ignored_exception_types` to `OpenAiImageGenerationDriver` by @vachillo in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1947\n### 📦 Dependencies\n* deps: bump the production-dependencies group with 8 updates by @dependabot[bot] in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1939\n* deps: bump the production-dependencies group with 10 updates by @dependabot[bot] in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1945\n### Other Changes\n* fix: doc string update for gpt-image-1 in OpenAiImageGenerationDriver by @ian-griptape-ai in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1936\n* ci: clean up stable release pipeline by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1937\n* Add AudioUrlArtifact by @zachgiordano in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1938\n* Document how to create ZIP-based tools and structures in Griptape Cloud by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1951\n* Fix headers for create structure\u002Ftool by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1953\n* fix(drivers-prompt-google): correctly handle debug logs by @collindutter in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1952\n* Support Griptape Cloud base_url with paths by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1954\n* chore: bump to next minor version in main branch by @vachillo in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1958\n* feat: Support gpt-image-1 in GriptapeCloudImageGenerationDriver by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1964\n* Handle embedded text with commas for PGAI Vector Store Driver by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1967\n* feat: MCP Tool by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1962\n* Remove dependency on active event_loop in MCP tool initialization by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1969\n* Docs\u002Fmcp by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1972\n* Super __attrs_post_init__ by @cjkindel in https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fpull\u002F1974\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.7.0...v1.8.0","2025-07-25T20:56:06",{"id":243,"version":244,"summary_zh":245,"released_at":246},135717,"v1.7.5","\u003C!-- Release notes generated using configuration in .github\u002Frelease.yml at v1.7.5 -->\n\n\n\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fgriptape-ai\u002Fgriptape\u002Fcompare\u002Fv1.7.4...v1.7.5","2025-07-08T19:32:21"]