--- title: "沟通与协作参考：amazon chime sdk meeting summarizer - 沟通与协作 | 思行致远" canonical: "https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-amazon-chime-sdk-meeting-summarizer/" dateModified: "2026-06-30" datePublished: "2026-06-30" --- 打开 AI·外脑目录 AI·外脑 [返回专栏首页](https://siyuanailab.com/knowledge/) × 站内搜索 ⌕ 主要栏目 - [AI 核心概念](https://siyuanailab.com/knowledge/ai-core/) - [工作流](https://siyuanailab.com/knowledge/workflows/) - [工具配置与选型](https://siyuanailab.com/knowledge/toolkits/) - [信源与标准](https://siyuanailab.com/knowledge/sources/) - [场景案例 待开发](https://siyuanailab.com/knowledge/scenarios/) - [职业与能力](https://siyuanailab.com/knowledge/career/) 专题 - [从这里开始](https://siyuanailab.com/knowledge/start/) - [获取与安装](https://siyuanailab.com/knowledge/install/) - [建设日志](https://siyuanailab.com/knowledge/ops/) [工作流子目录](https://siyuanailab.com/knowledge/workflows/) 基础设施层 - [浏览器自动化 **5**](https://siyuanailab.com/knowledge/workflows/infrastructure/browser-automation/) - [MCP 工具 **5**](https://siyuanailab.com/knowledge/workflows/infrastructure/mcp-tools/) - [文件系统与本地执行 **5**](https://siyuanailab.com/knowledge/workflows/infrastructure/filesystem-local/) - [API 网关与工具路由 **5**](https://siyuanailab.com/knowledge/workflows/infrastructure/api-tool-routing/) Skill 生产线 - [目标明确 **5**](https://siyuanailab.com/knowledge/workflows/skill-forge/goal-definition/) - [Skill 设计 **5**](https://siyuanailab.com/knowledge/workflows/skill-forge/skill-design/) - [代码实现 **5**](https://siyuanailab.com/knowledge/workflows/skill-forge/implementation/) - [测试评估 **5**](https://siyuanailab.com/knowledge/workflows/skill-forge/testing-evaluation/) Skill 参考库 - [写作与内容 **5**](https://siyuanailab.com/knowledge/workflows/skill-library/writing-content/) - [研究与知识 **5**](https://siyuanailab.com/knowledge/workflows/skill-library/research-knowledge/) - [沟通与协作 **5**](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/) - [WF-051 沟通与协作参考：ai meeting assistant](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-ai-meeting-assistant/) - [WF-052 沟通与协作参考：AI Meeting Notes Action Items Automator](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-ai-meeting-notes-action-items-automator/) - [WF-053 沟通与协作参考：AI Powered Meeting Summarizer](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-ai-powered-meeting-summarizer/) - [WF-054 沟通与协作参考：amazon chime sdk meeting summarizer](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-amazon-chime-sdk-meeting-summarizer/) - [WF-055 沟通与协作参考：MeetingSummarizer](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-meetingsummarizer/) - [产品与项目 **5**](https://siyuanailab.com/knowledge/workflows/skill-library/product-project/) - [数据与分析 **5**](https://siyuanailab.com/knowledge/workflows/skill-library/data-analysis/) - [创意与设计 **5**](https://siyuanailab.com/knowledge/workflows/skill-library/creative-design/) - [商业与营销 **5**](https://siyuanailab.com/knowledge/workflows/skill-library/business-marketing/) Agent Building 只收录有多步工作流、工作标准、Loop、Harness、工具链或通信机制的真 Agent；不收纯 persona、profile、curated list 或只有 system prompt 的仓库。 - [多 Agent 框架 **5**](https://siyuanailab.com/knowledge/workflows/agent-building/multi-agent-frameworks/) - [工作流编排 **5**](https://siyuanailab.com/knowledge/workflows/agent-building/workflow-orchestration/) - [Sub-agent 模式 **5**](https://siyuanailab.com/knowledge/workflows/agent-building/sub-agent-patterns/) - [Harness 与评估 **5**](https://siyuanailab.com/knowledge/workflows/agent-building/harness-evaluation/) Skill 参考库 · 沟通与协作 · Skill 解读 · WF-054 # 沟通与协作参考：amazon chime sdk meeting summarizer 思行致远 · 2026/6/30 · 置信度：低 会议纪要行动项协作沟通 一个 AWS CDK 方案，部署跨平台会议 bot，结合 Chime SDK、Transcribe、Bedrock 和知识库生成转写、摘要、回放和聊天查询。 # amazon chime sdk meeting summarizer > 一个 AWS CDK 方案，部署跨平台会议 bot，结合 Chime SDK、Transcribe、Bedrock 和知识库生成转写、摘要、回放和聊天查询。 ## 这个能力解决什么问题 它解决的是企业会议平台上的自动参会、转写、摘要和会后检索问题。输入是 Amazon Chime、Webex 或 Zoom 会议音频；输出包括 speaker-specific transcript、会议摘要、音频回放、知识库索引和 chatbot 查询能力。 ## 核心逻辑 用户通过前端安排会议后，CDK 部署的 Chime SDK/Voice 资源让 bot 加入会议并捕获音频；Amazon Transcribe 做转写和 diarization；Bedrock 模型生成摘要；S3/OpenSearch Serverless/Bedrock Knowledge Base 等资源保存材料并支持会后问答。部署、配置和站点资源由 Projen + AWS CDK 脚本管理。 ## 技术结构 - 关键模块：CDK stack、Chime SDK resources、Lambda handlers、S3 site bucket、Bedrock runtime/agent、Transcribe、OpenSearch Serverless、Scheduler、前端站点。 - 调用链路：schedule meeting -> bot joins/captures -> Transcribe -> Bedrock summarization -> storage/knowledge base -> frontend/chatbot。 - 输入输出：输入是会议音频和 AWS 配置；输出是 transcript、summary、audio playback、chatbot response。 - 核心依赖：AWS CDK v2、Projen、TypeScript、Amazon Chime SDK、Transcribe、Bedrock Anthropic/Titan、OpenSearch Serverless、S3、CloudFront。 ## 为什么值得参考 它展示了“会议智能”在云上落地需要哪些基础设施：音频捕获、转写、摘要、存储、检索、站点和权限资源，不只是一个 summarize prompt。README 开头的 AWS 内部禁用免责声明也提醒这类能力天然高风险。 ## 为什么不建议直接套用 项目明确声明未获 AWS 员工内部使用批准。自动加入会议、捕获音频、转写和保存回放都需要法律授权、参会者告知、数据驻留和访问审计。部署成本和 AWS 权限复杂，`yarn launch` 会创建大量云资源，不适合无云治理的个人或小团队。 ## 如何改造成自己的版本 不要从自动参会 bot 开始。先做“用户上传已授权 transcript/audio -> 摘要 -> 内部检索”的离线版。若必须接会议平台，先写 consent flow、数据保留策略、IAM 边界和资源销毁流程；所有会议资料默认私有，摘要发布必须人工确认。 ## 适用场景 - 有 AWS 能力和合规流程的企业会议智能试点。 - 研究会议 bot 云架构和 CDK 部署。 - 需要 transcript + summary + chatbot 的完整参考架构。 ## 不适用场景 - 未授权会议录音或自动入会。 - 没有 AWS 安全治理、预算和资源清理流程。 - 内部政策禁止生成式 AI 处理会议内容的组织。 ## 公开版边界 - 录音和转写须获得参会者知情同意。 ## 参考信息 - 原项目：[amazon-chime-sdk-meeting-summarizer](https://github.com/aws-samples/amazon-chime-sdk-meeting-summarizer) - 作者：aws-samples - 相关概念：[[会议智能]]、[[云端合规]] - 相关卡片：保守留空 ## 参考对象信息 **名称：**amazon-chime-sdk-meeting-summarizer **作者：**aws-samples **链接：**[https://github.com/aws-samples/amazon-chime-sdk-meeting-summarizer](https://github.com/aws-samples/amazon-chime-sdk-meeting-summarizer) ## 适用场景与行动 **场景：**Skill 参考库 / 沟通与协作 **适合谁：**关注 AI 工作流、Skill 生产和 Agent Building 的读者 **下一步：**先阅读边界说明，再判断它是否适合改造成自己的沟通与协作流程。 当前内容按公开来源和已审核草稿整理，只作为工作流理解与方法论参考。 涉及客户数据、外部平台、会议记录和业务系统的场景，必须保留人工确认与授权边界。 [上一张：沟通与协作参考：AI Powered Meeting Summarizer](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-ai-powered-meeting-summarizer/) [下一张：沟通与协作参考：MeetingSummarizer](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/skill-library-communication-meetingsummarizer/) [返回当前分类](https://siyuanailab.com/knowledge/workflows/skill-library/communication-collaboration/) [返回工作流专栏](https://siyuanailab.com/knowledge/workflows/)