跳到主要内容Skip to main content
投机取巧 Speculative Speculation

AI 工具AI Tools · Frameworks

框架与运行时Frameworks & runtimes

Agent 运行时 / 编排框架 / SDK 的责任边界、状态持久化能力与权限模型。收录标准：提供 Agent 运行时或编排层、有官方文档可回溯、近 30 天有实质更新。The responsibility boundaries of agent runtimes, orchestration frameworks and SDKs, plus state persistence and permission models. Admission criteria: provides an agent runtime or orchestration layer, has official documentation that can be traced back, and has substantive updates within the last 30 days.

30 个对象objects 8 个固定维度fixed dimensions 0 个已完整核验fully verified 最后核验Last checked 2026-11-03
  1. 本站按 10 大类收录 AI 产品。编程、对话、图像、视频、音频、研究、自动化、框架、MCP 工具、本地运行器——每类都有明确的收录标准，写在各大类的目录页里。This site files AI products into 10 categories. Coding, chat, image, video, audio, research, automation, frameworks, MCP tools and local runners — each with explicit admission criteria, written on the category page itself.
  2. 大类看目录，形态看字段。一个产品属于哪一类由收录边界决定；它怎么装进你的工作流（IDE / CLI / 厂商云 / 框架 / MCP…）是另一根独立的轴。Category comes from the shelf, form from a field. Which category a product belongs to is decided by admission boundaries; how it installs into your workflow (IDE / CLI / vendor cloud / framework / MCP…) is a separate axis.
  3. 每条档案都是核验快照。字段逐条挂官方出处与核验日期，本站不实测、不排名，未知就写未知。Every entry is a verification snapshot. Each field carries an official source and a checked date; we run no benchmarks, publish no rankings, and write "unknown" when it is unknown.

先回答一件事：你要不要自己搭Answer one question first: do you need to build one at all

这是本大类存在的唯一理由。如果现成产品已经够用，正确答案是「不要搭」——那就去看 AI 编程 那一类，不用往下看。真要自己搭，第一步是认清你要的抽象层。This is the only reason this section exists. If a finished product already does the job, the right answer is "do not build" — go to AI coding and stop reading. If you really are building, the first step is being clear which abstraction layer you want.

  1. 1我要最细的控制权，文件系统我自己管I want the finest control and will manage the filesystem myself
    原语型（primitives-only）Primitives only只给你 agent loop 的原语，不替你决定文件系统 / 规划 / 记忆怎么做 —— OpenAI Agents SDK 是这一类。You get the raw agent-loop primitives and it does not decide for you how the filesystem, planning or memory work — OpenAI Agents SDK is in this group.
  2. 2我要一个成熟 CLI 直接变成 API 对象I want a mature CLI turned into an API object
    包装既有 CLIWrapping an existing CLIClaude Agent SDK 捆绑完整 CLI，默认给全工具无沙箱 —— 你省了造轮子，但权限模型是继承来的。Claude Agent SDK bundles a complete CLI with all tools and no sandbox — you save reinventing it, but the permission model is inherited.
  3. 3我要开箱即跑的通用长任务助手I want a general long-task assistant that works out of the box
    batteries-includedBatteries includedDeep Agents 官方自述 opinionated：预置更多东西，也替你做了更多决定。Deep Agents describes itself as opinionated: more is preset, and more has been decided for you.
  4. 4流程要图、要能中断恢复、要多 Agent 协作I need a graph, interruptible resumption, and multi-agent collaboration
    编排框架：LangGraph / Google ADK / CrewAIOrchestration frameworks: LangGraph / Google ADK / CrewAILangGraph 的落盘粒度最明确（每 superstep 存一次图状态），ADK 的图执行引擎最完整，CrewAI 用「角色分工 + 事件流程」两套抽象。LangGraph has the clearest checkpoint granularity (graph state saved every superstep), ADK has the most complete graph execution engine, CrewAI uses two abstractions: role division and event flow.
  5. 5我要它常驻、跨平台、能定时跑I want it resident, cross-platform, able to run on a schedule
    平台型 / 常驻型：OpenHands / Hermes AgentPlatform / resident type: OpenHands / Hermes AgentOpenHands 是带 Web UI 的自托管控制中心；Hermes 更进一步——能挂 Telegram/Discord/Slack/WhatsApp/Signal/Email 六平台，还能 cron 无人值守。OpenHands is a self-hosted control centre with a web UI; Hermes goes further — it attaches to Telegram, Discord, Slack, WhatsApp, Signal and Email, and can run unattended on cron.
  6. 6我只想要个工具用，不想维护I just want a tool to use, not something to maintain
    不该来这个大类You should not be in this section这属于「不自建」那一侧，见 AI 编程。That belongs on the "do not build" side — see AI coding.

⚠ 三个最容易踩的坑：「状态」和「上下文」不是一回事——压缩上下文会丢信息，状态持久化才是长期可靠运行的前提（各家在这两层的完备程度差别很大）；「provider 无关」不等于「本地模型可用」（Claude Agent SDK 只支持 Claude，CrewAI 则点名支持 Ollama）；编排框架不给沙箱——权限边界要自己确认。⚠ Three traps people fall into most easily: "state" and "context" are not the same thing — compressing context loses information, while state persistence is what makes long-term reliable operation possible (and vendors differ a lot in how complete these two layers are). "provider-agnostic" does not mean "local models work" — Claude Agent SDK supports only Claude, while CrewAI explicitly names Ollama support. Orchestration frameworks provide no sandbox — you have to establish the permission boundary yourself.

按抽象层分组By abstraction layer

同一层里也能差很远 —— 两个「编程底座」一个是纯原语、一个是全托管包装 CLI，所以先按抽象层分，再看具体对象。Even within a layer the gap is wide — of two "coding bases", one is pure primitives and one is a fully managed CLI wrapper. So group by abstraction first, then look at the object.

编程底座Coding base 6

编程底座Coding base

给你 agent loop 或现成 CLI 的编程接口。这层的差别不在功能多少，而在它替你做多少决定 —— 一个只给原语，一个把整套 CLI 连权限模型一起打包给你。A programming interface giving you an agent loop or an existing CLI. The difference here is not how much it does but how many decisions it makes for you — one hands you primitives, the other packages the whole CLI along with its permission model.

Claude Agent SDKClaude Agent SDK

Anthropic

它是什么：Claude Code 的编程接口（Python 版）。 它把一个成熟的 coding CLI 变成 PythonThe Python programming interface to Claude Code — a polished c

部分核验Partly verified 详情 →Details →

Codex SDKCodex SDK

OpenAI

「把 Codex CLI 变成 Node 进程里可驱动的对象」。官方 README 首句就是 「Embed the Code"Turns the Codex CLI into an object you can drive inside a Nod

部分核验Partly verified 详情 →Details →

Open InterpreterOpen Interpreter

Open Interpreter（开源社区）Open Interpreter (open-source community)

Open Interpreter 把「开放模型 + 本地终端执行」做成产品（Apache-2.0）： 一个跑在你终端里的编码Open Interpreter pairs open models with local execution: a cod

部分核验Partly verified 详情 →Details →

Pydantic AIPydantic AI

Pydantic

Pydantic AI 是 Pydantic 出的 Python agent 库。 主打类型安全。官方自述把两个关键词并在一Pydantic AI is the type-safe Python agent loop — official READ

部分核验Partly verified 详情 →Details →

smolagentssmolagents

Hugging Face

smolagents 是一个薄 agent 库 —— 它让 agent「用代码思考」。 官方自述是 "barebones lA thin agent library from Hugging Face — it lets agents think

部分核验Partly verified 详情 →Details →

编排框架Orchestration 17

编排框架Orchestration

把多步骤、多 Agent、要中断恢复的流程显式建成图。给的是控制力，代价是接线与状态管理都归你。Makes multi-step, multi-agent flows with resumable state explicit graphs. It gives you control; the cost is that wiring and state management are yours.

AG2AG2

ag2ai

AG2 是 AutoGen 停更后由社区接棒的续作。 仓库描述直接写着 "AG2 (formerly AutoGen)"。 The community continuation of AutoGen after the original repo

部分核验Partly verified 详情 →Details →

AgentScopeAgentScope

阿里巴巴（魔搭社区）Alibaba (ModelScope)

AgentScope 是国产多智能体框架的主力线（Apache-2.0，阿里/魔搭系）：Msg 消息传递抽象清晰， 带可视化AgentScope is the flagship domestic multi-agent framework (Apa

部分核验Partly verified 详情 →Details →

CrewAICrewAI

CrewAI Inc

CrewAI 是一个 Python 框架。它用「角色分工」和「业务流程」两套比喻来编排多 agent。 官方自述是 "Fra"Organises multi-agent collaboration through role division and

部分核验Partly verified 详情 →Details →

DSPyDSPy

Stanford NLP / Databricks

DSPy 把「调提示词」变成编程问题（MIT，Stanford NLP/Databricks）：签名声明输入输出， 优化器把DSPy turns 'tuning prompts' into a programming problem (MIT, S

部分核验Partly verified 详情 →Details →

HaystackHaystack

deepset

Haystack 是「检索质量工程」那一派的 RAG 框架（Apache-2.0，deepset）：索引/查询管线分离、 hHaystack is the RAG framework of the 'retrieval-quality engine

部分核验Partly verified 详情 →Details →

LangChainLangChain

LangChain, Inc.

LangChain 是 LLM 应用框架生态的主体（MIT，148k★）：模型/工具/检索的统一抽象层， 提供多类集成入口。LangChain is the main body of the LLM-app framework ecosystem

部分核验Partly verified 详情 →Details →

LangChain4jLangChain4j

LangChain4j 社区LangChain4j community

LangChain4j 是 Java 生态的 LLM 应用框架：统一模型与向量库接口，支持工具调用与 MCP， 可嵌进 QuA Java LLM application framework: unified model and vector-sto

部分核验Partly verified 详情 →Details →

LangGraphLangGraph

LangChain Inc

LangGraph 是一个低层编排框架，专门用来构建有状态的 agent。 官方自述很克制，注意那个 low-level： LangGraph is a low-level orchestration framework for building

部分核验Partly verified 详情 →Details →

LlamaIndex FrameworkLlamaIndex Framework

LlamaIndex（run-llama）

一句话说清：LlamaIndex 是一个 Python 库。它把外部数据接进来做检索问答。 ⚠ 但有件事必须先说。 官方已明"A Python library that pulls external data in for retrieval an

部分核验Partly verified 详情 →Details →

MastraMastra

Mastra AI

Mastra 是「TypeScript 生态的全栈式 agent 框架」。 装一个包，Agents、图工作流、Memory、Mastra is the batteries-included agent framework for TypeScrip

部分核验Partly verified 详情 →Details →

MetaGPTMetaGPT

DeepWisdom（深度赋智，国产）DeepWisdom (China-origin)

MetaGPT 把「软件公司」写进多 agent 协作（MIT）：产品经理→架构师→工程师→QA 的角色流水线， SOP 编MetaGPT writes a "software company" into multi-agent collabora

部分核验Partly verified 详情 →Details →

Spring AISpring AI

Spring（VMware/Broadcom）Spring (VMware/Broadcom)

Spring AI 把 AI 能力做成 Spring 组件：可移植 Model API、ChatClient、向量库与 MCAI as Spring components: a portable Model API, ChatClient, vec

部分核验Partly verified 详情 →Details →

Vercel AI SDK 7Vercel AI SDK 7

VercelVercel

AI SDK 7 是 TypeScript/JS 的 agent 应用工具箱：统一的模型调用、工具调用与流式输出， 新增 HA TypeScript/JS agent toolkit: unified model and tool calls pl

部分核验Partly verified 详情 →Details →

通用 HarnessGeneral harness 7

通用 HarnessGeneral harness

预置了规划、文件系统、记忆等一整套，面向通用长任务。开箱程度最高，可改性最低。Presets planning, filesystem, memory and the rest, aimed at general long-running tasks. Most ready out of the box, least modifiable.

Deep AgentsDeep Agents

LangChain

Deep Agents 是 LangChain 出的通用 agent harness。 它把 filesystem、子代理、Deep Agents is a general-purpose agent harness from LangChain.

部分核验Partly verified 详情 →Details →

ElizaOS（Eliza）ElizaOS（Eliza）

ElizaOS 社区（Eliza Research）ElizaOS community (Eliza Research)

ElizaOS 把「角色」做成 agent 单元（MIT，TypeScript）：一个 character 配置文件定义人格ElizaOS makes the "character" the agent unit: one config file

部分核验Partly verified 详情 →Details →

Hermes AgentHermes Agent

Nous Research

Hermes Agent 是一个常驻的跨平台个人助手，会从经验里自己攒技能。 它不绑在你的笔记本上。 官方 README 说A resident personal assistant with a learning loop — put plain

部分核验Partly verified 详情 →Details →

Letta / MemGPTLetta / MemGPT

Letta, Inc.

Letta（原 MemGPT）把 agent 做成持久实体（Apache-2.0）：agent 服务器常驻、记忆分层 （coLetta (formerly MemGPT) makes agents persistent entities (Apac

部分核验Partly verified 详情 →Details →

SWE-agentSWE-agent

Princeton / Stanford（学术团队）Princeton / Stanford (academic teams)

SWE-agent 是「agent 修代码」研究线的参考实现（MIT，Princeton/Stanford）： ACI（为 SWE-agent is a research harness for agent-driven code repair (

部分核验Partly verified 详情 →Details →

本站目前是「核验快照」，还没有实测数据This site is currently a verification snapshot — no measured runs yet

下面每一个字段都能回到官方原文核对，但我们还没有在同一批任务上跑过这些工具。Every field below can be traced back to the official source, but we have not yet run these tools on the same batch of tasks. 实测协议已经写好（统一任务、统一验收、记录人工介入与返工次数），任务清单也已就绪，The measurement protocol is written (shared tasks, shared acceptance criteria, recording human interventions and rework counts) and the task list is ready, but it is 尚未执行。not yet run.

所以本站给的是能力边界与证据，不是「哪个更好用」的结论。So what this site gives you is capability boundaries and evidence, not a verdict on which is better to use. 真正的选型请用你自己的输入、预算与验收标准跑一遍。For an actual choice, run them yourself against your own inputs, budget and acceptance criteria.

8 个维度，每格都能回溯8 dimensions, every cell traceable

所有对象在同一坐标系下被描述。All objects are described on the same coordinate system. 点开任一对象，每一格下面都有官方源链接与核验日期——不认同可以自己回去查。Open any object and every cell carries an official source link and a checked date — disagree and verify it yourself.

01

模型与开放条件Model & openness

用谁的模型？能不能换第三方？多模型路由怎么配？Whose model does it use? Can you switch to a third party? How is multi-model routing configured?

02

运行位置Where it runs

本地 / 云端 / 混合？索引在本地还是云端？远程控制不等于云端执行。Local, cloud or hybrid? Is the index local or remote? Remote control is not the same as cloud execution.

03

本地文件Local files

能访问多大范围？未打开的大文件怎么读？符号级能力如何？How wide an area can it reach? How does it read large files you have not opened? How capable is it at the symbol level?

04

关机后的任务Tasks after shutdown

电脑关了还能继续跑吗？这是本地形态最常见的误解来源。Does it keep running once the computer is off? This is the most common misconception about local forms.

05

工具与扩展Tools & extensions

工具链有多宽：MCP · LSP · 插件 · Hooks · 自定义指令。How wide is the toolchain: MCP · LSP · plugins · hooks · custom instructions?

06

上下文与记忆Context & memory

索引算法、上下文窗口、会话恢复与分叉、跨会话记忆边界。Indexing algorithm, context window, session resume and fork, and the boundary of cross-session memory.

07

权限与限制Permissions & limits

沙箱与审批边界、凭据管理、数据是否出境、额度如何计算。Sandbox and approval boundaries, credential handling, whether data leaves your machine, and how quota is charged.

08

适合什么任务Best for

最擅长什么任务？什么情况下别用它？What tasks is it best at? When should you not use it?

全部数据一览All data at a glance

同一张表横向对比。点对象名进详情页看逐格证据。The same table side by side. Click an object name to see per-cell evidence.

对象Object形态Form厂商Vendor适合Best for核验日Checked状态Status
AGAGAG2AG2 框架 / SDKFramework / SDK ag2ai 适合。 要用 Python 编排多 agent 会话 / 协议化网络（Hub + channel，含咨询、讨论、工作流适配器） 需要可审计的Orchestrating multi-agent conversations / a protocol-driven network in 2026-10-23 partial
AgAGAgentScopeAgentScope 框架 / SDKFramework / SDK 阿里巴巴（魔搭社区）Alibaba (ModelScope) 适合：多智能体协作应用（消息传递模型清晰）、通义系栈、要可视化调试工作台。 不适合：单 agent 薄应用（多智能体抽象是成本）；要全网最大Fits: multi-agent collaboration apps, Tongyi-line stacks, Studio visua 2026-10-30 partial
AgAGAgno（原 Phidata）Agno (formerly Phidata) 框架 / SDKFramework / SDK Agno Inc. 适合：要「框架+控制平面」一体（跑起来+观测+调度）、Python 生态、要高频发版的活跃选型。 不适合：只要薄库不要平面、TS 生态（PyFits: framework + control plane as one (run/observe/schedule), Python 2026-10-31 partial
CACLClaude Agent SDKClaude Agent SDK 框架 / SDKFramework / SDK Anthropic 一句话：它是一个已经被打磨过的 coding agent。 换模型和自定义 loop 形状，它做不到。 适合： 想要一个已经被打磨过的 coOne line: it is a coding agent that has already been polished. Swappin 2026-10-23 partial
CXCOCodex SDKCodex SDK 框架 / SDKFramework / SDK OpenAI 适合四类情况： 要把 Codex 塞进自己的 Node 应用 / Electron 应用 / CI 需要非交互执行 需要精确到文件级的权限规A good fit in four cases: You want to put Codex inside your own Node a 2026-10-23 partial
CACRCrewAICrewAI 框架 / SDKFramework / SDK CrewAI Inc 适合三类场景。 任务能被表述成「一支有角色分工的团队」（Crews） 需要业务逻辑留在普通 Python 里、并显式控制执行路径（FlowsFits: tasks that can be expressed as "a team with defined roles" (Crew 2026-10-23 partial
DADEDeep AgentsDeep Agents 框架 / SDKFramework / SDK LangChain 适合三类场景。 要一个开箱即用、默认调优好、面向长任务多步流程的通用 Agent。 不想自己拼 filesystem / 记忆 / 子代理。Fits: you want a general agent that works out of the box, is tuned by 2026-10-23 partial
DSDSDSPyDSPy 框架 / SDKFramework / SDK Stanford NLP / Databricks 适合：把「调提示词」变成可复现的优化流程（有指标、有训练集）、多模型可换的管线。 不适合：要开箱 agent 运行时；小样本没指标时（优化器Fits: reproducible prompt-optimization flows, multi-model swappable pi 2026-10-30 partial
ElELElizaOS（Eliza）ElizaOS（Eliza） 框架 / SDKFramework / SDK ElizaOS 社区（Eliza Research）ElizaOS community (Eliza Research) 适合：要「有人格的常驻 agent」（社交/陪伴/社区 bot）、TS 生态、插件式扩展。 不适合：要编程任务底座、要企业工作流编排、对第三Fits: resident persona agents (social/community bots), the TS ecosyste 2026-10-31 partial
GAGOGoogle ADK（Agent Development Kit）Google ADK（Agent Development Kit） 框架 / SDKFramework / SDK Google 适合这些情况。 要把多个 agent 组织成有分支、有循环、有人工介入的确定性流程 要用代码优先（code-first）定义 agent 与Fits: organising multiple agents into deterministic flows with branche 2026-10-23 partial
HaHAHaystackHaystack 框架 / SDKFramework / SDK deepset 适合：生产级 RAG 管线（检索质量工程：hybrid/reranker/评估环）、要可审计的组件图。 不适合：只要一个薄聊天壳（过重）；aFits: production RAG (retrieval quality and auditable component graphs 2026-10-30 partial
HAHEHermes AgentHermes Agent 框架 / SDKFramework / SDK Nous Research 适合这六类场景。 想要一个常驻的跨平台个人助手 （六个平台：Telegram / Discord / Slack / WhatsApp / A fit for these six cases. A resident cross-platform personal assistan 2026-10-23 partial
LaLALangChainLangChain 框架 / SDKFramework / SDK LangChain, Inc. 适合：要最大集成生态、要「换 provider 不改代码」的抽象层、RAG/工具流起步。 不适合：要轻量直连单家 API（抽象层是成本）；要Fits: the largest provider/tool ecosystem, swapping models without rew 2026-10-30 partial
LaLALangChain4jLangChain4j 框架 / SDKFramework / SDK LangChain4j 社区LangChain4j community 适合：Java/Kotlin 团队、已有 Spring/Quarkus 服务、要把 LLM 与向量检索接进业务代码。 不适合：Python/Fits Java/Kotlin teams with existing Spring/Quarkus services who want 2026-11-03 partial
LGLALangGraphLangGraph 框架 / SDKFramework / SDK LangChain Inc 适合四类场景。 agent 流程要跑很久，且必须能从中断处精确恢复 需要人工在流程中间介入并修改状态 要把「工作记忆」与「跨会话长期记忆」分Fits: agent flows that run long and must resume precisely from an inte 2026-10-23 partial
LeLELetta / MemGPTLetta / MemGPT 框架 / SDKFramework / SDK Letta, Inc. 适合：要「持久人格」的 agent（长期记忆+状态续接）、要自托管 agent 服务器、MemGPT 设计路线。 不适合：无状态一次性调用（Fits: long-memory 'persistent persona' agents, self-hosted agent serve 2026-10-30 partial
LILLLlamaIndex FrameworkLlamaIndex Framework 框架 / SDKFramework / SDK LlamaIndex（run-llama） 适合：把外部数据接进来做检索问答。 核心任务是把外部数据源（API / PDF / 文档 / SQL）接进来做检索与问答 需要 300+ 现Fits: bringing external data sources in for retrieval and question ans 2026-10-23 partial
MAMAMastraMastra 框架 / SDKFramework / SDK Mastra AI 适合： TypeScript / Node 团队要一套「开箱即用」的 agent 框架 —— Agents + 图 Workflows + Fits: TypeScript / Node teams that want an out-of-the-box agent framew 2026-10-23 partial
MeMEMetaGPTMetaGPT 框架 / SDKFramework / SDK DeepWisdom（深度赋智，国产）DeepWisdom (China-origin) 适合：多智能体协作研究、SOP 流水线实验、软件生成流程的学术复现（论文引用量大）。 不适合：要活跃维护的生产选型（推送/发版放缓）、要精细Fits: multi-agent research, SOP pipeline experiments, paper-line repro 2026-10-31 partial
MAMIMicrosoft Agent FrameworkMicrosoft Agent Framework 框架 / SDKFramework / SDK Microsoft 它是「图工作流」式的编排框架，适合五类场景。 要把多 agent 组织成有分支、有人工介入、可 checkpoint 恢复的确定性图工作流 It is a graph-workflow orchestration framework — a fit for five cases. 2026-10-23 partial
OpOPOpen InterpreterOpen Interpreter 框架 / SDKFramework / SDK Open Interpreter（开源社区）Open Interpreter (open-source community) 适合：开放权重模型 + 本地终端编码 agent、要代码执行落在自己机器上、Rust 重写线的跟进者。 不适合：要云端托管执行、要企业管控面Fits: open-weight models + a local terminal coding agent, execution on 2026-10-31 partial
OAOPOpenAI Agents SDKOpenAI Agents SDK 框架 / SDKFramework / SDK OpenAI 适合： 想要「薄 harness」，agent loop 原语之外的自己挑。 需要 guardrails 与 human-in-the-loFits: you want a thin harness with built-in guardrails / human-in-the- 2026-10-23 partial
OHOPOpenHands Agent CanvasOpenHands Agent Canvas 框架 / SDKFramework / SDK OpenHands（All-Hands-AI） 能做的事： 把 agent 放到自己服务器上 7×24 跑 定时或 webhook 触发（Slack / GitHub / Linear /What it is good for: Running agents 7×24 on your own server Scheduled 2026-10-23 partial
PAPYPydantic AIPydantic AI 框架 / SDKFramework / SDK Pydantic 适合：需要类型安全的 Python 团队，以及要把 agent run 变成 durable 的人。 Python 团队要做「类型安全的结构A fit when: you are a Python team that wants type-safe structured outp 2026-10-23 partial
SeSESemantic Kernel（SK）Semantic Kernel（SK） 框架 / SDKFramework / SDK Microsoft 适合：.NET/企业栈要把 AI 嵌进现有应用、Azure 生态、要双语言（C#/Python）SDK。 不适合：新项目起步——官方已指明 Fits: .NET/enterprise stacks embedding AI into existing apps, the Azur 2026-10-31 partial
SMSMsmolagentssmolagents 框架 / SDKFramework / SDK Hugging Face 适合这四类场景。 认同 code-as-action（官方称比 JSON 工具调用少 30% 步数），并愿意给 LLM 生成代码上沙箱 要 Fits: teams that buy into code-as-action (officially ~30% fewer steps 2026-10-23 partial
SpSPSpring AISpring AI 框架 / SDKFramework / SDK Spring（VMware/Broadcom）Spring (VMware/Broadcom) 适合：Spring Boot 团队、要把 AI 能力做成标准 Spring 组件、需要 MCP 与向量库集成。 不适合：非 Java 团队、Fits Spring Boot teams who want AI as standard Spring components and n 2026-11-03 partial
StSTStrands Agents SDKStrands Agents SDK 框架 / SDKFramework / SDK AWS 适合：AWS 栈的 agent 开发（Bedrock 一等公民）、要显式 MCP 客户端能力、要轻量 SDK 不要重框架。 不适合：非 AWFits: AWS/Bedrock stacks, explicit MCP capability needs, lightweight c 2026-10-30 partial
SWSWSWE-agentSWE-agent 框架 / SDKFramework / SDK Princeton / Stanford（学术团队）Princeton / Stanford (academic teams) 适合：研究/评测「agent 修代码」能力（SWE-bench 线）、要 ACI 设计参考实现。 不适合：生产编码助手开箱（它是研究向 haFits: evaluating/researching coding agents, ACI design references, SWE 2026-10-30 partial
VeVEVercel AI SDK 7Vercel AI SDK 7 框架 / SDKFramework / SDK VercelVercel 适合：TypeScript/JS 团队自建 agent 应用，想在不改业务代码的前提下切换模型或 harness。 不适合：要开箱即用的成品Fits TypeScript/JS teams building agent apps who want to switch models 2026-11-03 partial

为什么这里不给总分排名Why there is no overall ranking

这不是省略，是方法上的明确取舍。This is not an omission but a deliberate methodological choice.

  • 缺统一实测，就不给分。跨工具的跑分口径不同——题目、预算、执行环境都不一样，A 的 90 分和 B 的 88 分不可比。No shared measurement, no score. Benchmarks across tools use different rules — different questions, budgets and environments. A 90 and a B 88 are not comparable.
  • 「未知」是合法答案。查不到就写「未知」并说明为什么，不用推测填充。缺信息本身也是信息。"Unknown" is a legitimate answer. If it cannot be found, we write "unknown" and explain why, instead of filling the gap with guesses. Missing information is itself information.
  • 来源链接、类型与核验日都在详情页里，可以逐条回去复核。The source link, its kind and the checked date are all on the detail page, so you can walk any of them back and re-check.
  • lifecycle 单独标注。核验日新不等于数据有效——上游归档后数据会失效，所以单独标 active / maintenance / archived。lifecycle is marked separately. A recent check date does not mean the data is still valid — it goes stale when upstream archives the repo, so active / maintenance / archived are tracked separately.
  • 不同族不混排。找现成的、自己搭的、接上与跑起来的是三种角色，坐标系不同；硬合成一张总表只会制造假的可比性。Families are not mixed. Ready-made tools, things you build yourself, and plug-in-and-run components are three different roles with different coordinate systems; forcing them into one master table would manufacture false comparability.

可信度标记怎么读How to read the confidence marks

详情页与上表都带Both the detail pages and the table above carry confidence，它反映, which reflects 本站数据的完整度how complete this site's data is，不是对产品的评价。, not a judgement of the product.

verified8 个维度均有官方源支撑，且核验日在 90 天内all 8 dimensions have official sources and were checked within 90 days
partial部分维度标为未知，或官方文档不可访问，或核验日超过 90 天some dimensions are marked unknown, or the official docs were unreachable, or checked over 90 days ago
stale官方已发布重大变化，本站尚未核验the vendor has shipped a significant change that this site has not re-checked

本站当前的完整度Current completeness on this site

数据层用Recheck this figure with node scripts/audit-gaps.mjs 可随时复核这个数字。. 下表实时计算（本页档案最后核验日 2026-11-03）：Computed from current data; profiles last checked 2026-11-03:

大类Category已补齐Filled说明Meaning
框架与运行时Frameworks & runtimes 179/240 编排框架的版本演进快、接口变化多，逐条核验成本最高 —— 也是全站最低的一档Orchestration frameworks move fast and change APIs often, so these are the most expensive to verify — and the lowest fill rate on the site

全站合计 1372/1896 维度已补齐Site-wide 1372/1896 dimensions filled（72%）

剩余未补齐项分三类：官方未公开（索引算法、沙箱实现细节本就不对外说明）、需实测才能确定（大仓库表现、CI 无 TTY 行为）、客观渠道不可达。The remaining gaps fall into three kinds: not disclosed by the vendor (indexing algorithms, sandbox internals are simply not documented), requires measurement (behaviour on very large repositories, CI without a TTY), and unreachable through objective channels.

我们选择留白而不是填「已支持」——错误的成本最终由使用者承担。We chose to leave a blank rather than fill in "supported" — the cost of an error is ultimately borne by whoever uses it.

数据来源与复核方式Sources and how to re-check them

全部数据来自开源仓库All data comes from the open-source repository speculcom/ai-handbook（CC BY 4.0）。 (CC BY 4.0).

每个条目都是纯 Markdown + frontmatter，含维度记录、证据链接、核验日与可信度标记。仓库内含：Every entry is plain Markdown plus frontmatter, carrying dimension records and an evidence chain with links, checked dates and confidence marks. The repository contains:

  • tools/METHODOLOGY.md —— 方法论总纲与核心规则the methodology charter and core rules
  • tools/axes/ —— 每个维度的定义与判定标准（含正反例）the definition and decision criteria for each dimension (with positive and negative examples)
  • tools/SCHEMA.md —— 数据格式规范与构建校验规则the data-format spec and build validation rules
  • tools/protocols/<大类><category>/_protocol.md —— 实测协议(the measurement protocol (尚未执行not yet run，故本站暂无实测结论), so this site has no measured conclusions)

发现错误或有新证据，欢迎提 Issue 或 PR——每条修正都会注明依据与影响范围。Found an error or have new evidence? Open an issue or a PR — each correction notes its basis and blast radius.