# Docs by LangChain

> Documentation for LangSmith, Fleet, and our open source packages.

## Docs

- [Create Connection](https://cndoc-langchain.site/api-reference/agent-connections-v2/create-connection.md)
- [List Connections](https://cndoc-langchain.site/api-reference/agent-connections-v2/list-connections.md)
- [Remove Connection](https://cndoc-langchain.site/api-reference/agent-connections-v2/remove-connection.md)
- [Authenticate](https://cndoc-langchain.site/api-reference/auth-service-v2/authenticate.md): Get OAuth token or start authentication flow if needed.
- [Check Oauth Token Exists](https://cndoc-langchain.site/api-reference/auth-service-v2/check-oauth-token-exists.md): Return whether the current user has any tokens for a given provider (across agents).
- [Check Workspace Slack Tokens Exist](https://cndoc-langchain.site/api-reference/auth-service-v2/check-workspace-slack-tokens-exist.md): Check if the workspace has any Slack tokens.
- [Create Mcp Oauth Provider](https://cndoc-langchain.site/api-reference/auth-service-v2/create-mcp-oauth-provider.md): Create an OAuth provider via MCP auto-discovery.
- [Create Oauth Provider](https://cndoc-langchain.site/api-reference/auth-service-v2/create-oauth-provider.md): Create a new OAuth provider manually.
- [Delete Oauth Provider](https://cndoc-langchain.site/api-reference/auth-service-v2/delete-oauth-provider.md): Delete an OAuth provider.
- [Delete Oauth Tokens For User](https://cndoc-langchain.site/api-reference/auth-service-v2/delete-oauth-tokens-for-user.md): Delete all tokens for the current user for the given provider (across agents).
- [Delete Single Oauth Token](https://cndoc-langchain.site/api-reference/auth-service-v2/delete-single-oauth-token.md): Delete a specific OAuth token, revoking it at the provider first.
- [Get Oauth Provider](https://cndoc-langchain.site/api-reference/auth-service-v2/get-oauth-provider.md): Get a specific OAuth provider.
- [List Oauth Providers](https://cndoc-langchain.site/api-reference/auth-service-v2/list-oauth-providers.md): List OAuth providers.
- [List Oauth Tokens For User](https://cndoc-langchain.site/api-reference/auth-service-v2/list-oauth-tokens-for-user.md): List the calling user's tokens for a provider.
- [Oauth Callback](https://cndoc-langchain.site/api-reference/auth-service-v2/oauth-callback.md)
- [Oauth Callback Get](https://cndoc-langchain.site/api-reference/auth-service-v2/oauth-callback-get.md): Handle OAuth callback redirect from OAuth providers.
- [Oauth Setup Callback](https://cndoc-langchain.site/api-reference/auth-service-v2/oauth-setup-callback.md): Handle OAuth setup callback redirect from GitHub Apps.
- [Revoke All Slack Tokens For Workspace](https://cndoc-langchain.site/api-reference/auth-service-v2/revoke-all-slack-tokens-for-workspace.md): Revoke ALL Slack tokens for the workspace. Admin-only action that disconnects Slack entirely.
- [Update Oauth Provider](https://cndoc-langchain.site/api-reference/auth-service-v2/update-oauth-provider.md): Update an OAuth provider.
- [Update Token Label](https://cndoc-langchain.site/api-reference/auth-service-v2/update-token-label.md): Update a token's provider_account_label. Only the token owner can update.
- [Wait For Auth Completion](https://cndoc-langchain.site/api-reference/auth-service-v2/wait-for-auth-completion.md): Wait for OAuth authentication completion.
- [Create Deployment](https://cndoc-langchain.site/api-reference/deployments-v2/create-deployment.md): Create a new deployment.
- [Delete Deployment](https://cndoc-langchain.site/api-reference/deployments-v2/delete-deployment.md): Delete a deployment by ID.
- [Delete Deployments](https://cndoc-langchain.site/api-reference/deployments-v2/delete-deployments.md): Delete multiple deployments with partial success support.
- [Get Deployment](https://cndoc-langchain.site/api-reference/deployments-v2/get-deployment.md): Get a deployment by ID.
- [Get Revision](https://cndoc-langchain.site/api-reference/deployments-v2/get-revision.md): Get a revision by ID for a deployment.
- [List Deployments](https://cndoc-langchain.site/api-reference/deployments-v2/list-deployments.md): List all deployments.
- [List Revisions](https://cndoc-langchain.site/api-reference/deployments-v2/list-revisions.md): List all revisions for a deployment.
- [Patch Deployment](https://cndoc-langchain.site/api-reference/deployments-v2/patch-deployment.md): Patch a deployment by ID.
- [Redeploy Revision](https://cndoc-langchain.site/api-reference/deployments-v2/redeploy-revision.md): Redeploy a specific revision ID.
- [List Forge GitHub Integrations](https://cndoc-langchain.site/api-reference/integrations-v1/list-forge-github-integrations.md): List available Forge GitHub integrations.
- [List Forge GitHub Repositories](https://cndoc-langchain.site/api-reference/integrations-v1/list-forge-github-repositories.md): List available GitHub repositories for a Forge integration.
- [List GitHub Integrations](https://cndoc-langchain.site/api-reference/integrations-v1/list-github-integrations.md): List available GitHub integrations for LangGraph Platfom Cloud SaaS.
- [List GitHub Repositories](https://cndoc-langchain.site/api-reference/integrations-v1/list-github-repositories.md): List available GitHub repositories for an integration that are available to deploy to LangSmith Deployment.
- [Create Listener](https://cndoc-langchain.site/api-reference/listeners-v2/create-listener.md): Create a listener.<br> <br> Creating a listener is only allowed for LangSmith organizations with self-hosted enterprise plans.
- [Delete Listener](https://cndoc-langchain.site/api-reference/listeners-v2/delete-listener.md): Delete a listener by ID.
- [Get Listener](https://cndoc-langchain.site/api-reference/listeners-v2/get-listener.md): Get a listener by ID.
- [List Listeners](https://cndoc-langchain.site/api-reference/listeners-v2/list-listeners.md): List all listeners.
- [Patch Listener](https://cndoc-langchain.site/api-reference/listeners-v2/patch-listener.md): Patch a listener by ID.
- [](https://cndoc-langchain.site/index.md)
- [基于属性的访问控制](https://cndoc-langchain.site/langsmith/abac.md)
- [在被追踪的函数中访问当前运行（跨度）](https://cndoc-langchain.site/langsmith/access-current-span.md)
- [连接身份验证提供者](https://cndoc-langchain.site/langsmith/add-auth-server.md)
- [使用服务器 API 实现人机交互](https://cndoc-langchain.site/langsmith/add-human-in-the-loop.md)
- [为跟踪添加元数据和标签](https://cndoc-langchain.site/langsmith/add-metadata-tags.md)
- [概览](https://cndoc-langchain.site/langsmith/administration-overview.md)
- [设置代理认证（Beta 版）](https://cndoc-langchain.site/langsmith/agent-auth.md): 使用代理认证，通过 OAuth 2.0 凭证实现从代理到任何系统的安全访问。
- [Agent Server](https://cndoc-langchain.site/langsmith/agent-server.md)
- [A2A JSON-RPC](https://cndoc-langchain.site/langsmith/agent-server-api/a2a/a2a-json-rpc.md): Communicate with an assistant using the Agent-to-Agent (A2A) Protocol over JSON-RPC 2.0. This endpoint accepts a JSON-RPC envelope and dispatches based on `method`.
- [Count Assistants](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/count-assistants.md): Get the count of assistants matching the specified criteria.
- [Create Assistant](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/create-assistant.md): Create an assistant.
- [Delete Assistant](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/delete-assistant.md): Delete an assistant by ID.
- [Get Assistant](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/get-assistant.md): Get an assistant by ID.
- [Get Assistant Graph](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/get-assistant-graph.md): Get an assistant by ID.
- [Get Assistant Schemas](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/get-assistant-schemas.md): Get an assistant by ID.
- [Get Assistant Subgraphs](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/get-assistant-subgraphs.md): Get an assistant's subgraphs.
- [Get Assistant Subgraphs by Namespace](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/get-assistant-subgraphs-by-namespace.md): Get an assistant's subgraphs filtered by namespace.
- [Get Assistant Versions](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/get-assistant-versions.md): Get all versions of an assistant.
- [Patch Assistant](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/patch-assistant.md): Update an assistant.
- [Search Assistants](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/search-assistants.md): Search for assistants.
- [Set Latest Assistant Version](https://cndoc-langchain.site/langsmith/agent-server-api/assistants/set-latest-assistant-version.md): Set the latest version for an assistant.
- [Count Crons](https://cndoc-langchain.site/langsmith/agent-server-api/crons/count-crons.md): Get the count of crons matching the specified criteria.
- [Create Cron](https://cndoc-langchain.site/langsmith/agent-server-api/crons/create-cron.md): Create a cron to schedule runs on new threads.
- [Create Thread Cron](https://cndoc-langchain.site/langsmith/agent-server-api/crons/create-thread-cron.md): Create a cron to schedule runs on a thread.
- [Delete Cron](https://cndoc-langchain.site/langsmith/agent-server-api/crons/delete-cron.md): Delete a cron by ID.
- [Search Crons](https://cndoc-langchain.site/langsmith/agent-server-api/crons/search-crons.md): Search all active crons
- [Update Cron](https://cndoc-langchain.site/langsmith/agent-server-api/crons/update-cron.md): Update a cron job by ID.
- [MCP Get](https://cndoc-langchain.site/langsmith/agent-server-api/mcp/mcp-get.md): Implemented according to the Streamable HTTP Transport specification.
- [MCP Post](https://cndoc-langchain.site/langsmith/agent-server-api/mcp/mcp-post.md): Implemented according to the Streamable HTTP Transport specification. Sends a JSON-RPC 2.0 message to the server.
- [Terminate Session](https://cndoc-langchain.site/langsmith/agent-server-api/mcp/terminate-session.md): Implemented according to the Streamable HTTP Transport specification. Terminate an MCP session. The server implementation is stateless, so this is a no-op.
- [Create Background Run](https://cndoc-langchain.site/langsmith/agent-server-api/stateless-runs/create-background-run.md): Create a run and return the run ID immediately. Don't wait for the final run output.
- [Create Run Batch](https://cndoc-langchain.site/langsmith/agent-server-api/stateless-runs/create-run-batch.md): Create a batch of runs and return immediately.
- [Create Run, Stream Output](https://cndoc-langchain.site/langsmith/agent-server-api/stateless-runs/create-run-stream-output.md): Create a run and stream the output.
- [Create Run, Wait for Output](https://cndoc-langchain.site/langsmith/agent-server-api/stateless-runs/create-run-wait-for-output.md): Create a run, wait for the final output and then return it.
- [Delete an item.](https://cndoc-langchain.site/langsmith/agent-server-api/store/delete-an-item.md)
- [List namespaces with optional match conditions.](https://cndoc-langchain.site/langsmith/agent-server-api/store/list-namespaces-with-optional-match-conditions.md)
- [Retrieve a single item.](https://cndoc-langchain.site/langsmith/agent-server-api/store/retrieve-a-single-item.md)
- [Search or list items within a namespace prefix.](https://cndoc-langchain.site/langsmith/agent-server-api/store/search-or-list-items-within-a-namespace-prefix.md): Lists items ordered by last updated time. If a `query` is provided, performs a natural language search instead. Supports pagination via `limit` and `offset`, and filtering via `filter`.
- [Store or update an item.](https://cndoc-langchain.site/langsmith/agent-server-api/store/store-or-update-an-item.md)
- [API Documentation](https://cndoc-langchain.site/langsmith/agent-server-api/system/api-documentation.md): A local reference to the Agent Server API documentation.
- [Health Check](https://cndoc-langchain.site/langsmith/agent-server-api/system/health-check.md): Check the health status of the server. Optionally check database connectivity.
- [Server Information](https://cndoc-langchain.site/langsmith/agent-server-api/system/server-information.md): Get server version information, feature flags, and metadata.
- [System Metrics](https://cndoc-langchain.site/langsmith/agent-server-api/system/system-metrics.md): Get system metrics in Prometheus or JSON format for monitoring and observability.
- [Cancel Run](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/cancel-run.md)
- [Cancel Runs](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/cancel-runs.md): Cancel one or more runs. Can cancel runs by thread ID and run IDs, or by status filter.
- [Create Background Run](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/create-background-run.md): Create a run in existing thread, return the run ID immediately. Don't wait for the final run output.
- [Create Run, Stream Output](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/create-run-stream-output.md): Create a run in existing thread. Stream the output.
- [Create Run, Wait for Output](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/create-run-wait-for-output.md): Create a run in existing thread. Wait for the final output and then return it.
- [Delete Run](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/delete-run.md): Delete a run by ID.
- [Get Run](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/get-run.md): Get a run by ID.
- [Join Run](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/join-run.md): Wait for a run to finish.
- [Join Run Stream](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/join-run-stream.md): Join a run stream. This endpoint streams output in real-time from a run similar to the /threads/__THREAD_ID__/runs/stream endpoint. If the run has been created with `stream_resumable=true`, the stream can be resumed from the last seen event ID.
- [List Runs](https://cndoc-langchain.site/langsmith/agent-server-api/thread-runs/list-runs.md): List runs for a thread.
- [Copy Thread](https://cndoc-langchain.site/langsmith/agent-server-api/threads/copy-thread.md): Create a new thread with a copy of the state and checkpoints from an existing thread.
- [Count Threads](https://cndoc-langchain.site/langsmith/agent-server-api/threads/count-threads.md): Get the count of threads matching the specified criteria.
- [Create Thread](https://cndoc-langchain.site/langsmith/agent-server-api/threads/create-thread.md): Create a thread.
- [Delete Thread](https://cndoc-langchain.site/langsmith/agent-server-api/threads/delete-thread.md): Delete a thread by ID.
- [Get Thread](https://cndoc-langchain.site/langsmith/agent-server-api/threads/get-thread.md): Get a thread by ID.
- [Get Thread History](https://cndoc-langchain.site/langsmith/agent-server-api/threads/get-thread-history.md): Get all past states for a thread.
- [Get Thread History Post](https://cndoc-langchain.site/langsmith/agent-server-api/threads/get-thread-history-post.md): Get all past states for a thread.
- [Get Thread State](https://cndoc-langchain.site/langsmith/agent-server-api/threads/get-thread-state.md): Get state for a thread.
- [Get Thread State At Checkpoint](https://cndoc-langchain.site/langsmith/agent-server-api/threads/get-thread-state-at-checkpoint.md): Get state for a thread at a specific checkpoint.
- [Get Thread State At Checkpoint](https://cndoc-langchain.site/langsmith/agent-server-api/threads/get-thread-state-at-checkpoint-1.md): Get state for a thread at a specific checkpoint.
- [Join Thread Stream](https://cndoc-langchain.site/langsmith/agent-server-api/threads/join-thread-stream.md): This endpoint streams output in real-time from a thread. The stream will include the output of each run executed sequentially on the thread and will remain open indefinitely. It is the responsibility of the calling client to close the connection.
- [Patch Thread](https://cndoc-langchain.site/langsmith/agent-server-api/threads/patch-thread.md): Update a thread.
- [Prune Threads](https://cndoc-langchain.site/langsmith/agent-server-api/threads/prune-threads.md): Prune threads by ID. The 'delete' strategy removes threads entirely. The 'keep_latest' strategy prunes old checkpoints but keeps threads and their latest state.
- [Search Threads](https://cndoc-langchain.site/langsmith/agent-server-api/threads/search-threads.md): Search for threads.
- [Update Thread State](https://cndoc-langchain.site/langsmith/agent-server-api/threads/update-thread-state.md): Add state to a thread.
- [Agent Server 更新日志](https://cndoc-langchain.site/langsmith/agent-server-changelog.md)
- [使用代理服务器进行分布式追踪](https://cndoc-langchain.site/langsmith/agent-server-distributed-tracing.md): 当从另一个服务调用您部署的代理服务器时，使用 RemoteGraph 或 SDK 统一追踪。
- [如何为 Agent Server 运行收集用户反馈](https://cndoc-langchain.site/langsmith/agent-server-feedback.md)
- [为扩展配置 LangSmith 代理服务器](https://cndoc-langchain.site/langsmith/agent-server-scale.md)
- [LangSmith 中的警报](https://cndoc-langchain.site/langsmith/alerts.md)
- [分析实验](https://cndoc-langchain.site/langsmith/analyze-an-experiment.md)
- [自定义检测](https://cndoc-langchain.site/langsmith/annotate-code.md): 直接检测您的代码，以控制哪些函数被追踪以及它们在 LangSmith 中的显示方式。
- [内联标注跟踪和运行](https://cndoc-langchain.site/langsmith/annotate-traces-inline.md)
- [使用标注队列](https://cndoc-langchain.site/langsmith/annotation-queues.md)
- [通过编程方式管理反馈与标注队列](https://cndoc-langchain.site/langsmith/annotation-queues-sdk.md)
- [LangSmith 部署的控制平面 API 参考](https://cndoc-langchain.site/langsmith/api-ref-control-plane.md)
- [应用结构](https://cndoc-langchain.site/langsmith/application-structure.md)
- [助手](https://cndoc-langchain.site/langsmith/assistants.md)
- [使用 SDK 记录用户反馈](https://cndoc-langchain.site/langsmith/attach-user-feedback.md)
- [如何审核评估器评分](https://cndoc-langchain.site/langsmith/audit-evaluator-scores.md)
- [审计日志](https://cndoc-langchain.site/langsmith/audit-logs.md): 跟踪和审查您整个 LangSmith 组织中的管理操作，以实现安全性、合规性和运营可见性。
- [认证与访问控制](https://cndoc-langchain.site/langsmith/auth.md)
- [认证方法](https://cndoc-langchain.site/langsmith/authentication-methods.md)
- [在 AWS 上自托管 LangSmith](https://cndoc-langchain.site/langsmith/aws-self-hosted.md)
- [在 Azure 上自托管 LangSmith](https://cndoc-langchain.site/langsmith/azure-self-hosted.md)
- [如何启动后台运行](https://cndoc-langchain.site/langsmith/background-run.md)
- [在您的账户中管理账单](https://cndoc-langchain.site/langsmith/billing.md)
- [在实验上自动运行评估器](https://cndoc-langchain.site/langsmith/bind-evaluator-to-dataset.md)
- [使用服务器端缓存](https://cndoc-langchain.site/langsmith/caching.md): 在您的智能体部署中，使用 stale-while-revalidate 和键值缓存 API 在服务器端缓存值。
- [如何取消运行](https://cndoc-langchain.site/langsmith/cancel-run.md): 通过 API 取消单个或多个运行，并选择中断或回滚操作。
- [使用 LangSmith 部署和评估实现 CI/CD 管道](https://cndoc-langchain.site/langsmith/cicd-pipeline-example.md)
- [LangGraph CLI](https://cndoc-langchain.site/langsmith/cli.md)
- [云](https://cndoc-langchain.site/langsmith/cloud.md)
- [如何定义代码评估器](https://cndoc-langchain.site/langsmith/code-evaluator-sdk.md)
- [如何定义代码评估器](https://cndoc-langchain.site/langsmith/code-evaluator-ui.md)
- [如何比较实验结果](https://cndoc-langchain.site/langsmith/compare-experiment-results.md)
- [LangSmith 部署组件](https://cndoc-langchain.site/langsmith/components.md): 代理服务器、LangGraph CLI、Studio、SDK、RemoteGraph、控制平面和数据平面组件的概述。
- [如何创建复合评估器](https://cndoc-langchain.site/langsmith/composite-evaluators-sdk.md)
- [如何创建复合评估器](https://cndoc-langchain.site/langsmith/composite-evaluators-ui.md)
- [条件追踪](https://cndoc-langchain.site/langsmith/conditional-tracing.md)
- [使用 HTTP 头进行运行时配置](https://cndoc-langchain.site/langsmith/configurable-headers.md)
- [在服务器日志中包含HTTP头](https://cndoc-langchain.site/langsmith/configurable-logs.md)
- [管理助手](https://cndoc-langchain.site/langsmith/configuration-cloud.md)
- [配置检查点存储后端](https://cndoc-langchain.site/langsmith/configure-checkpointer.md): 配置 Agent Server 以使用 PostgreSQL、MongoDB 或自定义实现来存储检查点。
- [配置运行输入和输出预览](https://cndoc-langchain.site/langsmith/configure-input-output-preview.md): 通过为特定追踪类型配置自定义预览路径，自定义运行表中输入和输出列显示的内容。
- [如何为您的应用添加生存时间（TTL）](https://cndoc-langchain.site/langsmith/configure-ttl.md)
- [LangSmith 控制平面](https://cndoc-langchain.site/langsmith/control-plane.md)
- [核心能力概览](https://cndoc-langchain.site/langsmith/core-capabilities.md): Agent Server 核心能力概览，包括流式处理、人机协同、MCP、A2A、分布式追踪、Webhooks 和双重消息处理。
- [成本跟踪](https://cndoc-langchain.site/langsmith/cost-tracking.md)
- [创建提示词](https://cndoc-langchain.site/langsmith/create-a-prompt.md)
- [创建账户和 API 密钥](https://cndoc-langchain.site/langsmith/create-account-api-key.md)
- [如何使用少样本示例改进您的评估器](https://cndoc-langchain.site/langsmith/create-few-shot-evaluators.md)
- [使用定时任务](https://cndoc-langchain.site/langsmith/cron-jobs.md)
- [添加自定义认证](https://cndoc-langchain.site/langsmith/custom-auth.md)
- [如何使用自定义检查点器](https://cndoc-langchain.site/langsmith/custom-checkpointer.md): 在您的智能体部署中，使用自定义的 BaseCheckpointSaver 实现替换内置的 Postgres 检查点器。
- [如何自定义 Dockerfile](https://cndoc-langchain.site/langsmith/custom-docker.md)
- [连接自定义模型](https://cndoc-langchain.site/langsmith/custom-endpoint.md)
- [如何添加自定义生命周期事件](https://cndoc-langchain.site/langsmith/custom-lifespan.md)
- [如何添加自定义中间件](https://cndoc-langchain.site/langsmith/custom-middleware.md)
- [连接到 OpenAI 兼容的模型提供商/代理](https://cndoc-langchain.site/langsmith/custom-openai-compliant-model.md)
- [自定义输出渲染](https://cndoc-langchain.site/langsmith/custom-output-rendering.md)
- [如何添加自定义路由](https://cndoc-langchain.site/langsmith/custom-routes.md)
- [如何使用自定义存储](https://cndoc-langchain.site/langsmith/custom-store.md): 在您的智能体部署中，使用自定义的 BaseStore 实现替换内置的 Postgres 存储。
- [使用仪表板监控项目](https://cndoc-langchain.site/langsmith/dashboards.md)
- [批量导出追踪数据](https://cndoc-langchain.site/langsmith/data-export.md): 将 LangSmith 追踪数据以 Parquet 格式导出到兼容 S3 的存储桶。
- [管理批量导出目标](https://cndoc-langchain.site/langsmith/data-export-destinations.md): 为 LangSmith 批量导出配置和管理 S3 兼容的导出目标。
- [导入导出数据](https://cndoc-langchain.site/langsmith/data-export-downstream.md): 将 LangSmith 批量导出的 Parquet 数据导入 BigQuery、Snowflake、Redshift、Clickhouse 或 DuckDB。
- [监控和故障排除批量导出](https://cndoc-langchain.site/langsmith/data-export-monitor.md): 监控批量导出状态、管理正在运行的导出并排除故障。
- [LangSmith 数据平面](https://cndoc-langchain.site/langsmith/data-plane.md)
- [合规性数据清除](https://cndoc-langchain.site/langsmith/data-purging-compliance.md)
- [数据存储与隐私](https://cndoc-langchain.site/langsmith/data-storage-and-privacy.md)
- [数据集预置JSON模式类型](https://cndoc-langchain.site/langsmith/dataset-json-types.md)
- [数据集转换](https://cndoc-langchain.site/langsmith/dataset-transformations.md)
- [如何定义用于评估的目标函数](https://cndoc-langchain.site/langsmith/define-target-function.md)
- [设置混合部署 LangSmith](https://cndoc-langchain.site/langsmith/deploy-hybrid.md): 将自托管数据平面连接到托管的 LangSmith 控制平面，以实现混合代理部署。
- [部署其他框架](https://cndoc-langchain.site/langsmith/deploy-other-frameworks.md): 使用 LangGraph 函数式 API 将使用 Strands、CrewAI 或其他框架构建的智能体部署到 LangSmith。
- [启用 LangSmith 部署、Fleet、Insights 和 Polly](https://cndoc-langchain.site/langsmith/deploy-self-hosted-full-platform.md): 在自托管 LangSmith 实例上启用 LangSmith 部署、Fleet、Insights 和 Polly。
- [自托管独立服务器](https://cndoc-langchain.site/langsmith/deploy-standalone-server.md): 使用 Docker、Docker Compose 或 Kubernetes 部署独立的 Agent 服务器，无需 LangSmith 控制平面。
- [在云端部署](https://cndoc-langchain.site/langsmith/deploy-to-cloud.md): 创建和管理 LangSmith Cloud 部署，包括修订版、日志、指标和设置。
- [使用控制平面部署](https://cndoc-langchain.site/langsmith/deploy-with-control-plane.md): 使用控制平面 UI 构建 Docker 镜像并将应用程序部署到混合或自托管的 LangSmith 实例。
- [LangSmith 部署](https://cndoc-langchain.site/langsmith/deployment.md): 使用持久执行、实时流式传输和水平扩展来部署和管理智能体。
- [将您的应用部署到云端](https://cndoc-langchain.site/langsmith/deployment-quickstart.md): 使用 LangGraph CLI 将您的第一个应用部署到 LangSmith Cloud。
- [将您的深度智能体应用部署到云端](https://cndoc-langchain.site/langsmith/deployment-quickstart-da.md)
- [自托管部署的故障排除](https://cndoc-langchain.site/langsmith/diagnostics-self-hosted.md): 在联系支持团队之前，用于排查自托管 LangSmith 部署问题的诊断步骤。
- [实现分布式追踪](https://cndoc-langchain.site/langsmith/distributed-tracing.md)
- [双重文本](https://cndoc-langchain.site/langsmith/double-texting.md)
- [添加静态加密](https://cndoc-langchain.site/langsmith/encryption.md)
- [排队并发](https://cndoc-langchain.site/langsmith/enqueue-concurrent.md)
- [LangSmith 企业版](https://cndoc-langchain.site/langsmith/enterprise.md): 面向企业用户的部署选项、访问控制、数据隐私、成本控制和安全合规。
- [环境变量](https://cndoc-langchain.site/langsmith/env-var.md)
- [评估聊天机器人](https://cndoc-langchain.site/langsmith/evaluate-chatbot-tutorial.md)
- [评估复杂智能体](https://cndoc-langchain.site/langsmith/evaluate-complex-agent.md)
- [如何向现有实验添加评估器（仅限 Python）](https://cndoc-langchain.site/langsmith/evaluate-existing-experiment.md)
- [如何评估图](https://cndoc-langchain.site/langsmith/evaluate-graph.md)
- [如何评估LLM应用](https://cndoc-langchain.site/langsmith/evaluate-llm-application.md)
- [如何评估应用程序的中间步骤](https://cndoc-langchain.site/langsmith/evaluate-on-intermediate-steps.md)
- [如何运行成对评估](https://cndoc-langchain.site/langsmith/evaluate-pairwise.md)
- [评估 RAG 应用](https://cndoc-langchain.site/langsmith/evaluate-rag-tutorial.md)
- [使用多模态内容运行评估](https://cndoc-langchain.site/langsmith/evaluate-with-attachments.md): 了解如何创建带有文件附件的数据集示例，并在运行LangSmith多模态内容评估时，将其用于提示词和评估器。
- [如何使用 OpenTelemetry 进行评估](https://cndoc-langchain.site/langsmith/evaluate-with-opentelemetry.md)
- [如何在实验中重试失败的运行（仅限Python）](https://cndoc-langchain.site/langsmith/evaluate-with-retry.md)
- [LangSmith 评估](https://cndoc-langchain.site/langsmith/evaluation.md)
- [特定应用的评估方法](https://cndoc-langchain.site/langsmith/evaluation-approaches.md)
- [如何异步运行评估](https://cndoc-langchain.site/langsmith/evaluation-async.md)
- [评估概念](https://cndoc-langchain.site/langsmith/evaluation-concepts.md)
- [评估快速入门](https://cndoc-langchain.site/langsmith/evaluation-quickstart.md)
- [评估类型](https://cndoc-langchain.site/langsmith/evaluation-types.md)
- [管理评估器](https://cndoc-langchain.site/langsmith/evaluators.md): 在 LangSmith 中查看和管理工作区级别的评估器。
- [示例数据格式](https://cndoc-langchain.site/langsmith/example-data-format.md)
- [实验配置](https://cndoc-langchain.site/langsmith/experiment-configuration.md)
- [将 LangSmith 遥测数据导出到您的可观测性后端](https://cndoc-langchain.site/langsmith/export-backend.md)
- [使用 SDK 查询追踪](https://cndoc-langchain.site/langsmith/export-traces.md)
- [常见问题解答](https://cndoc-langchain.site/langsmith/faq.md)
- [反馈数据格式](https://cndoc-langchain.site/langsmith/feedback-data-format.md)
- [如何获取实验的性能指标](https://cndoc-langchain.site/langsmith/fetch-perf-metrics-experiment.md)
- [如何在UI中筛选实验](https://cndoc-langchain.site/langsmith/filter-experiments-ui.md)
- [过滤追踪](https://cndoc-langchain.site/langsmith/filter-traces-in-application.md)
- [访问与监督](https://cndoc-langchain.site/langsmith/fleet/access-and-oversight.md): 控制谁可以访问代理，他们如何进行身份验证，并审计他们的一切行为。
- [代理身份](https://cndoc-langchain.site/langsmith/fleet/agent-identity.md): 选择您的 Fleet 代理是使用自身凭据进行身份验证，还是使用每个用户的凭据进行身份验证。
- [Arcade 集成](https://cndoc-langchain.site/langsmith/fleet/arcade.md): 将您的工作区连接到 Arcade，使智能体能够访问 GitHub、Gmail、Slack 等第三方工具。
- [认证感知的工具响应](https://cndoc-langchain.site/langsmith/fleet/auth-format.md): 格式化工具响应以触发 OAuth 流程并自动恢复执行。
- [渠道](https://cndoc-langchain.site/langsmith/fleet/channels.md): 配置渠道以自动触发您的 Fleet 代理。
- [从代码调用代理](https://cndoc-langchain.site/langsmith/fleet/code.md): 通过 REST API 从 Python、JavaScript 或任何语言调用 Fleet 代理。
- [代理平台对比](https://cndoc-langchain.site/langsmith/fleet/comparison.md): 对比 LangSmith Fleet 与 Claude Cowork、Amazon Quick 和 Google Workspace Studio，为您的团队选择合适的企业代理平台
- [核心要素](https://cndoc-langchain.site/langsmith/fleet/essentials.md): Fleet 的核心功能
- [LangSmith Fleet](https://cndoc-langchain.site/langsmith/fleet/index.md): 无需代码即可创建有用的 AI 代理。从模板开始，连接您的账户，让代理处理日常工作，同时您保持控制权。
- [管理代理设置](https://cndoc-langchain.site/langsmith/fleet/manage-agent-settings.md): 在 Fleet 中管理您的代理。
- [LangSmith 工具服务器](https://cndoc-langchain.site/langsmith/fleet/mcp-framework.md)
- [定价常见问题](https://cndoc-langchain.site/langsmith/fleet/pricing-faq.md)
- [快速入门](https://cndoc-langchain.site/langsmith/fleet/quickstart.md): 从模板构建一个智能体
- [远程 MCP 服务器](https://cndoc-langchain.site/langsmith/fleet/remote-mcp-servers.md): 将 Fleet 连接到流行的远程 MCP 服务器
- [Salesforce 集成](https://cndoc-langchain.site/langsmith/fleet/salesforce.md): 将 LangSmith Fleet 连接到 Salesforce，以便您的代理可以查询记录、导航模式并读取自定义字段。
- [计划任务](https://cndoc-langchain.site/langsmith/fleet/schedules.md): 配置计划任务以定期运行您的 Fleet 代理。
- [自托管](https://cndoc-langchain.site/langsmith/fleet/self-hosted-link.md)
- [技能](https://cndoc-langchain.site/langsmith/fleet/skills.md): 使用技能为您的代理提供特定功能的访问权限。
- [将 Slack 与代理集成](https://cndoc-langchain.site/langsmith/fleet/slack-app.md): 将 LangSmith Fleet 连接到您的 Slack 工作区，让您的代理能够在 Slack 中与用户沟通。
- [将 Teams 与智能体集成](https://cndoc-langchain.site/langsmith/fleet/teams-app.md): 通过自带 Azure 机器人将 LangSmith Fleet 连接到 Microsoft Teams，让智能体能在 Teams 中与用户通信。
- [模板](https://cndoc-langchain.site/langsmith/fleet/templates.md): 通过精选的 Fleet 模板快速启动，并自定义工具、提示词和通道。
- [工具集成](https://cndoc-langchain.site/langsmith/fleet/tools.md): 为您的代理提供访问各种工具和服务的能力。
- [Fleet Webhooks](https://cndoc-langchain.site/langsmith/fleet/webhooks.md): 将智能体发布与外部系统、CI/CD 管道或自定义部署工作流集成。
- [管理工作区管理](https://cndoc-langchain.site/langsmith/fleet/workspace-admin.md): 为 Fleet 配置工作区级别的设置。
- [在 GCP 上自托管 LangSmith](https://cndoc-langchain.site/langsmith/gcp-self-hosted.md)
- [如何使用 LangGraph 实现生成式用户界面](https://cndoc-langchain.site/langsmith/generative-ui-react.md)
- [集成](https://cndoc-langchain.site/langsmith/get-started-integrations.md)
- [细粒度计费使用量](https://cndoc-langchain.site/langsmith/granular-usage.md): 获取按工作区、项目、用户或API密钥细分的详细跟踪使用数据。
- [在运行时重建图](https://cndoc-langchain.site/langsmith/graph-rebuild.md): 使用 ServerRuntime 为每次运行使用不同配置重建你的图。
- [LangSmith 文档](https://cndoc-langchain.site/langsmith/home.md)
- [使用服务器 API 进行时间旅行](https://cndoc-langchain.site/langsmith/human-in-the-loop-time-travel.md)
- [混合模式](https://cndoc-langchain.site/langsmith/hybrid.md)
- [使用人类反馈改进LLM作为评判者的评估器](https://cndoc-langchain.site/langsmith/improve-judge-evaluator-feedback.md)
- [使用 Insights Agent 发现错误与使用模式](https://cndoc-langchain.site/langsmith/insights.md)
- [集成](https://cndoc-langchain.site/langsmith/integrations.md)
- [中断并发](https://cndoc-langchain.site/langsmith/interrupt-concurrent.md)
- [在 SSO 组织中管理用户访问](https://cndoc-langchain.site/langsmith/jit-invite-sso.md)
- [在 Kubernetes 上自托管 LangSmith](https://cndoc-langchain.site/langsmith/kubernetes.md)
- [如何评估可运行对象](https://cndoc-langchain.site/langsmith/langchain-runnable.md)
- [LangGraph JS/TS SDK](https://cndoc-langchain.site/langsmith/langgraph-js-ts-sdk.md)
- [LangGraph Python SDK](https://cndoc-langchain.site/langsmith/langgraph-python-sdk.md)
- [LangSmith CLI](https://cndoc-langchain.site/langsmith/langsmith-cli.md): 从终端查询和管理 LangSmith 项目、追踪、运行、数据集、评估器、实验和线程
- [为 LangSmith 遥测配置您的收集器](https://cndoc-langchain.site/langsmith/langsmith-collector.md)
- [LangSmith 托管的 ClickHouse](https://cndoc-langchain.site/langsmith/langsmith-managed-clickhouse.md)
- [LangSmith MCP 服务器](https://cndoc-langchain.site/langsmith/langsmith-mcp-server.md): 使用模型上下文协议（MCP）服务器，让语言模型能够从 LangSmith 获取对话历史、提示词、运行记录、数据集、实验和账单信息。
- [如何定义 LLM-as-a-judge 评估器](https://cndoc-langchain.site/langsmith/llm-as-judge.md)
- [如何定义LLM作为评判者的评估器](https://cndoc-langchain.site/langsmith/llm-as-judge-sdk.md)
- [设置 LLM 认证代理](https://cndoc-langchain.site/langsmith/llm-auth-proxy-self-hosted.md): 部署一个基于 Envoy 的认证代理，用于验证 LangSmith 签名的 JWT 并将 LLM 请求路由到您的上游提供商或网关。
- [如何在本地运行评估（仅限Python）](https://cndoc-langchain.site/langsmith/local.md)
- [本地开发与测试](https://cndoc-langchain.site/langsmith/local-dev-testing.md): 比较 langgraph dev 和 langgraph up，用于代理服务器应用程序的本地开发和类生产环境测试。
- [记录 LLM 调用](https://cndoc-langchain.site/langsmith/log-llm-trace.md)
- [记录多模态追踪](https://cndoc-langchain.site/langsmith/log-multimodal-traces.md)
- [日志检索器追踪](https://cndoc-langchain.site/langsmith/log-retriever-trace.md): LangSmith 追踪中的日志检索步骤，用于文档级查看您的 RAG 管道。
- [将追踪记录到特定项目](https://cndoc-langchain.site/langsmith/log-traces-to-project.md): 使用环境变量或 SDK 将 LangSmith 追踪路由到指定项目，而非默认项目。
- [元数据参数参考](https://cndoc-langchain.site/langsmith/ls-metadata-parameters.md)
- [管理数据集](https://cndoc-langchain.site/langsmith/manage-datasets.md)
- [在 UI 中创建和管理数据集](https://cndoc-langchain.site/langsmith/manage-datasets-in-application.md)
- [如何通过SDK程序化创建和管理数据集](https://cndoc-langchain.site/langsmith/manage-datasets-programmatically.md)
- [使用 API 管理您的组织](https://cndoc-langchain.site/langsmith/manage-organization-by-api.md)
- [管理提示](https://cndoc-langchain.site/langsmith/manage-prompts.md): 在 LangSmith 中管理提示版本、环境和访问控制。
- [以编程方式管理提示](https://cndoc-langchain.site/langsmith/manage-prompts-programmatically.md)
- [管理追踪](https://cndoc-langchain.site/langsmith/manage-trace.md): 在 LangSmith 的追踪视图中并排比较追踪、公开共享追踪并查看服务器日志。
- [配置提示词设置](https://cndoc-langchain.site/langsmith/managing-model-configurations.md)
- [防止在跟踪中记录敏感数据](https://cndoc-langchain.site/langsmith/mask-inputs-outputs.md)
- [如何返回分类指标与数值指标](https://cndoc-langchain.site/langsmith/metric-type.md)
- [管理模型配置](https://cndoc-langchain.site/langsmith/model-configurations.md): 管理模型配置并控制其在 LangSmith 各功能中的可用性。
- [单体仓库支持](https://cndoc-langchain.site/langsmith/monorepo-support.md)
- [如何模拟多轮交互](https://cndoc-langchain.site/langsmith/multi-turn-simulation.md)
- [在提示中包含多模态内容](https://cndoc-langchain.site/langsmith/multimodal-content.md)
- [测试多轮对话](https://cndoc-langchain.site/langsmith/multiple-messages.md)
- [如何在单个评估器中返回多个分数](https://cndoc-langchain.site/langsmith/multiple-scores.md)
- [排查追踪嵌套问题](https://cndoc-langchain.site/langsmith/nest-traces.md)
- [LangSmith 可观察性](https://cndoc-langchain.site/langsmith/observability.md): 使用 LangSmith 为您的 LLM 应用程序添加检测、调查追踪并在生产环境中监控性能。
- [可观测性概念](https://cndoc-langchain.site/langsmith/observability-concepts.md)
- [追踪LLM应用教程](https://cndoc-langchain.site/langsmith/observability-llm-tutorial.md): 在LLM应用的原型设计、测试和生产阶段添加LangSmith可观测性。
- [追踪快速入门](https://cndoc-langchain.site/langsmith/observability-quickstart.md): 在几分钟内为大语言模型应用添加 LangSmith 追踪。
- [为您的 LangSmith 部署部署可观测性套件](https://cndoc-langchain.site/langsmith/observability-stack.md)
- [Studio 中的可观测性](https://cndoc-langchain.site/langsmith/observability-studio.md)
- [设置在线代码评估器](https://cndoc-langchain.site/langsmith/online-evaluations-code.md)
- [设置复合在线评估器](https://cndoc-langchain.site/langsmith/online-evaluations-composite.md)
- [设置 LLM-as-a-judge 在线评估器](https://cndoc-langchain.site/langsmith/online-evaluations-llm-as-judge.md)
- [设置多轮在线评估器](https://cndoc-langchain.site/langsmith/online-evaluations-multi-turn.md)
- [在 OpenAPI 中记录 API 认证](https://cndoc-langchain.site/langsmith/openapi-security.md)
- [使用 openevals 包运行评估](https://cndoc-langchain.site/langsmith/openevals.md): 使用开源的 openevals 和 agentevals 包与 LangSmith 一起运行评估。
- [优化分类器](https://cndoc-langchain.site/langsmith/optimize-classifier.md)
- [组织和工作区操作参考](https://cndoc-langchain.site/langsmith/organization-workspace-operations.md)
- [使用 OpenTelemetry 网关架构脱敏敏感数据](https://cndoc-langchain.site/langsmith/otel-gateway-trace-redaction.md): 使用 OpenTelemetry 收集器在追踪数据进入 LangSmith 之前脱敏敏感数据。
- [设置 LangSmith](https://cndoc-langchain.site/langsmith/platform-setup.md)
- [模型提供商](https://cndoc-langchain.site/langsmith/playground-model-providers.md)
- [LangSmith Polly](https://cndoc-langchain.site/langsmith/polly.md)
- [LangSmith Polly](https://cndoc-langchain.site/langsmith/polly-evaluation.md)
- [LangSmith Polly](https://cndoc-langchain.site/langsmith/polly-observability.md)
- [LangSmith Polly](https://cndoc-langchain.site/langsmith/polly-prompt-engineering.md)
- [使用预签名 URL 收集反馈](https://cndoc-langchain.site/langsmith/presigned-feedback-tokens.md): 使用预签名反馈令牌从客户端应用程序收集用户反馈，无需暴露您的 LangSmith API 密钥。
- [常见问题解答](https://cndoc-langchain.site/langsmith/pricing-faq.md)
- [定价方案](https://cndoc-langchain.site/langsmith/pricing-plans.md)
- [如何与 GitHub 同步提示词](https://cndoc-langchain.site/langsmith/prompt-commit.md)
- [提示工程](https://cndoc-langchain.site/langsmith/prompt-engineering.md)
- [提示工程概念](https://cndoc-langchain.site/langsmith/prompt-engineering-concepts.md)
- [提示工程快速入门](https://cndoc-langchain.site/langsmith/prompt-engineering-quickstart.md)
- [提示模板格式指南](https://cndoc-langchain.site/langsmith/prompt-template-format.md)
- [如何使用 pytest 运行评估](https://cndoc-langchain.site/langsmith/pytest.md)
- [使用 SDK 查询线程](https://cndoc-langchain.site/langsmith/query-threads.md): 从你的 LangSmith 项目中以编程方式获取和检查多轮对话线程。
- [Studio 快速入门](https://cndoc-langchain.site/langsmith/quick-start-studio.md)
- [如何处理模型速率限制](https://cndoc-langchain.site/langsmith/rate-limiting.md)
- [基于角色的访问控制](https://cndoc-langchain.site/langsmith/rbac.md)
- [如何在本地读取实验结果](https://cndoc-langchain.site/langsmith/read-local-experiment-results.md)
- [LangSmith 参考文档](https://cndoc-langchain.site/langsmith/reference.md)
- [区域常见问题](https://cndoc-langchain.site/langsmith/regions-faq.md)
- [拒绝并发](https://cndoc-langchain.site/langsmith/reject-concurrent.md)
- [发布策略](https://cndoc-langchain.site/langsmith/release-versions.md): 自托管 LangSmith 的发布渠道、节奏和版本支持。
- [RemoteGraph](https://cndoc-langchain.site/langsmith/remote-graph.md)
- [如何进行重复评估](https://cndoc-langchain.site/langsmith/repetition.md)
- [使对话私有化](https://cndoc-langchain.site/langsmith/resource-auth.md)
- [回滚并发](https://cndoc-langchain.site/langsmith/rollback-concurrent.md)
- [设置自动化规则](https://cndoc-langchain.site/langsmith/rules.md)
- [在智能体的新版本上运行回测](https://cndoc-langchain.site/langsmith/run-backtests-new-agent.md)
- [运行（跨度）数据格式](https://cndoc-langchain.site/langsmith/run-data-format.md)
- [如何使用 REST API](https://cndoc-langchain.site/langsmith/run-evals-api-only.md)
- [从演练场运行评估](https://cndoc-langchain.site/langsmith/run-evaluation-from-playground.md)
- [运行](https://cndoc-langchain.site/langsmith/runs.md): Agent Server 中运行的概述，包括如何启动后台运行、无状态运行以及取消运行。
- [如何在同一线程上运行多个代理](https://cndoc-langchain.site/langsmith/same-thread.md)
- [设置追踪的采样率](https://cndoc-langchain.site/langsmith/sample-traces.md)
- [沙箱认证代理](https://cndoc-langchain.site/langsmith/sandbox-auth-proxy.md): 在不硬编码密钥的情况下，将凭据注入来自沙箱的出站 API 请求。
- [沙箱 CLI](https://cndoc-langchain.site/langsmith/sandbox-cli.md): 通过命令行创建、检查、连接到 LangSmith 沙箱，并建立隧道。
- [沙盒 SDK 使用指南](https://cndoc-langchain.site/langsmith/sandbox-sdk.md): 使用 Python 或 TypeScript SDK 以编程方式创建和管理沙盒。
- [沙箱服务 URL](https://cndoc-langchain.site/langsmith/sandbox-service-urls.md): 通过经过身份验证的 URL，从浏览器或以编程方式访问在沙箱内运行的 HTTP 服务。
- [沙盒快照](https://cndoc-langchain.site/langsmith/sandbox-snapshots.md): 构建并捕获快照——每个沙盒启动时所依据的文件系统镜像。
- [沙盒概述](https://cndoc-langchain.site/langsmith/sandboxes.md): 使用托管沙盒在隔离环境中安全地执行代码并与文件系统交互。
- [可扩展性与弹性](https://cndoc-langchain.site/langsmith/scalability-and-resilience.md)
- [删除工作空间](https://cndoc-langchain.site/langsmith/script-delete-a-workspace.md)
- [删除组织](https://cndoc-langchain.site/langsmith/script-delete-an-organization.md)
- [删除追踪记录](https://cndoc-langchain.site/langsmith/script-delete-traces.md)
- [生成 ClickHouse 统计信息](https://cndoc-langchain.site/langsmith/script-generate-clickhouse-stats.md)
- [生成查询统计信息](https://cndoc-langchain.site/langsmith/script-generate-query-stats.md)
- [对 ClickHouse 运行支持查询](https://cndoc-langchain.site/langsmith/script-running-ch-support-queries.md)
- [针对 PostgreSQL 运行支持查询](https://cndoc-langchain.site/langsmith/script-running-pg-support-queries.md)
- [使用电子邮件和密码进行基础认证](https://cndoc-langchain.site/langsmith/self-host-basic-auth.md)
- [启用 Blob 存储](https://cndoc-langchain.site/langsmith/self-host-blob-storage.md)
- [配置自定义 TLS 证书](https://cndoc-langchain.site/langsmith/self-host-custom-tls-certificates.md)
- [用于计费和运营遥测的出站连接](https://cndoc-langchain.site/langsmith/self-host-egress.md)
- [连接到外部 ClickHouse 数据库](https://cndoc-langchain.site/langsmith/self-host-external-clickhouse.md)
- [连接到外部 PostgreSQL 数据库](https://cndoc-langchain.site/langsmith/self-host-external-postgres.md)
- [连接到外部 Redis 或 Valkey 数据库](https://cndoc-langchain.site/langsmith/self-host-external-redis.md)
- [FIPS 合规镜像](https://cndoc-langchain.site/langsmith/self-host-fips.md): 在 FIPS 140 合规的容器镜像上运行自托管 LangSmith 安装
- [为安装创建 Ingress（Kubernetes）](https://cndoc-langchain.site/langsmith/self-host-ingress.md)
- [为您的 LangSmith 安装创建镜像](https://cndoc-langchain.site/langsmith/self-host-mirroring-images.md)
- [查看整个组织的追踪计数](https://cndoc-langchain.site/langsmith/self-host-organization-charts.md)
- [为模型提供商使用环境变量](https://cndoc-langchain.site/langsmith/self-host-playground-environment-settings.md)
- [为扩展配置 LangSmith](https://cndoc-langchain.site/langsmith/self-host-scale.md)
- [使用 OAuth2.0 和 OIDC 设置 SSO](https://cndoc-langchain.site/langsmith/self-host-sso.md)
- [启用 TTL 和数据保留](https://cndoc-langchain.site/langsmith/self-host-ttl.md)
- [自定义错误支持消息](https://cndoc-langchain.site/langsmith/self-host-ui-customization.md): 为自托管部署在 LangSmith 前端中自定义支持联系信息。
- [升级安装](https://cndoc-langchain.site/langsmith/self-host-upgrades.md)
- [与您的自托管 LangSmith 实例交互](https://cndoc-langchain.site/langsmith/self-host-usage.md)
- [自定义用户管理](https://cndoc-langchain.site/langsmith/self-host-user-management.md)
- [为您的安装使用现有密钥（Kubernetes）](https://cndoc-langchain.site/langsmith/self-host-using-an-existing-secret.md)
- [自托管 LangSmith](https://cndoc-langchain.site/langsmith/self-hosted.md)
- [自托管 LangSmith 更新日志](https://cndoc-langchain.site/langsmith/self-hosted-changelog.md)
- [如何为您的代理部署添加语义搜索](https://cndoc-langchain.site/langsmith/semantic-search.md)
- [Agent Server 中的 A2A 端点](https://cndoc-langchain.site/langsmith/server-a2a.md): 使用 A2A 协议在 LangSmith 中实现支持分布式追踪的代理间通信。
- [LangSmith 部署的 Agent Server API 参考](https://cndoc-langchain.site/langsmith/server-api-ref.md)
- [代理服务器中的 MCP 端点](https://cndoc-langchain.site/langsmith/server-mcp.md)
- [在无服务器环境中追踪 JS 函数](https://cndoc-langchain.site/langsmith/serverless-environments.md)
- [设置自定义身份验证](https://cndoc-langchain.site/langsmith/set-up-custom-auth.md)
- [设置反馈标准](https://cndoc-langchain.site/langsmith/set-up-feedback-criteria.md)
- [设置层级结构](https://cndoc-langchain.site/langsmith/set-up-hierarchy.md)
- [设置资源标签](https://cndoc-langchain.site/langsmith/set-up-resource-tags.md): 创建和管理资源标签，以组织 LangSmith 工作区内的项目、数据集、提示词和其他资源。
- [如何使用 requirements.txt 设置应用程序](https://cndoc-langchain.site/langsmith/setup-app-requirements-txt.md)
- [如何设置 JavaScript 应用程序](https://cndoc-langchain.site/langsmith/setup-javascript.md)
- [如何使用 pyproject.toml 设置应用程序](https://cndoc-langchain.site/langsmith/setup-pyproject.md)
- [LangSmith 共享责任模型](https://cndoc-langchain.site/langsmith/shared-responsibility-model.md): 概述 LangChain 与客户如何共同承担 LangSmith 平台的安全责任。
- [LangSmith 技能](https://cndoc-langchain.site/langsmith/skills.md): 使用代理技能从您的编码代理中操作 LangSmith 追踪、数据集和评估器。
- [LangSmith API参考](https://cndoc-langchain.site/langsmith/smith-api-ref.md)
- [LangSmith 部署 SDK](https://cndoc-langchain.site/langsmith/smith-deployments-sdk.md)
- [LangSmith Go SDK](https://cndoc-langchain.site/langsmith/smith-go-sdk.md)
- [LangSmith Java SDK](https://cndoc-langchain.site/langsmith/smith-java-sdk.md)
- [LangSmith JS/TS SDK](https://cndoc-langchain.site/langsmith/smith-js-ts-sdk.md)
- [LangSmith Python SDK](https://cndoc-langchain.site/langsmith/smith-python-sdk.md)
- [无状态运行](https://cndoc-langchain.site/langsmith/stateless-runs.md)
- [LangSmith 状态](https://cndoc-langchain.site/langsmith/status.md)
- [流式 API](https://cndoc-langchain.site/langsmith/streaming.md)
- [LangSmith Studio](https://cndoc-langchain.site/langsmith/studio.md)
- [如何定义汇总评估器](https://cndoc-langchain.site/langsmith/summary.md)
- [使用 Pytest/Vitest 和 LangSmith 测试 ReAct 代理](https://cndoc-langchain.site/langsmith/test-react-agent-pytest.md)
- [配置线程](https://cndoc-langchain.site/langsmith/threads.md)
- [追踪 Anthropic 应用](https://cndoc-langchain.site/langsmith/trace-anthropic.md)
- [追踪 Amazon Bedrock 应用](https://cndoc-langchain.site/langsmith/trace-bedrock.md)
- [追踪 Claude Agent SDK 应用](https://cndoc-langchain.site/langsmith/trace-claude-agent-sdk.md)
- [追踪 Claude Code 应用](https://cndoc-langchain.site/langsmith/trace-claude-code.md)
- [追踪 Deep Agents 应用](https://cndoc-langchain.site/langsmith/trace-deep-agents.md)
- [追踪 DeepSeek 应用](https://cndoc-langchain.site/langsmith/trace-deepseek.md)
- [追踪生成器函数](https://cndoc-langchain.site/langsmith/trace-generator-functions.md)
- [追踪 LiteLLM 应用](https://cndoc-langchain.site/langsmith/trace-litellm.md)
- [追踪OpenAI应用](https://cndoc-langchain.site/langsmith/trace-openai.md)
- [追踪查询语法](https://cndoc-langchain.site/langsmith/trace-query-syntax.md)
- [使用 API 进行追踪](https://cndoc-langchain.site/langsmith/trace-with-api.md): 了解如何直接使用 LangSmith REST API 追踪 LLM 应用程序。
- [追踪 AutoGen 应用](https://cndoc-langchain.site/langsmith/trace-with-autogen.md)
- [追踪 CrewAI 应用](https://cndoc-langchain.site/langsmith/trace-with-crewai.md)
- [追踪 Google ADK 应用](https://cndoc-langchain.site/langsmith/trace-with-google-adk.md)
- [追踪 Google Gemini 应用](https://cndoc-langchain.site/langsmith/trace-with-google-gemini.md)
- [追踪 Instructor 应用](https://cndoc-langchain.site/langsmith/trace-with-instructor.md)
- [追踪 LangChain 应用（Python 和 JS/TS）](https://cndoc-langchain.site/langsmith/trace-with-langchain.md)
- [追踪 LangGraph 应用](https://cndoc-langchain.site/langsmith/trace-with-langgraph.md)
- [追踪 LiveKit 应用](https://cndoc-langchain.site/langsmith/trace-with-livekit.md)
- [追踪 Mastra 应用](https://cndoc-langchain.site/langsmith/trace-with-mastra.md)
- [追踪 Microsoft Agent Framework 应用程序](https://cndoc-langchain.site/langsmith/trace-with-microsoft-agent-framework.md)
- [追踪 Mistral 应用](https://cndoc-langchain.site/langsmith/trace-with-mistral.md)
- [追踪 n8n 工作流](https://cndoc-langchain.site/langsmith/trace-with-n8n.md): 了解如何在 LangSmith 中追踪 n8n AI 工作流。
- [追踪 OpenAI Agents SDK 应用程序](https://cndoc-langchain.site/langsmith/trace-with-openai-agents-sdk.md): 使用 LangSmith 追踪 OpenAI Agents SDK 的 Python 和 JavaScript 应用程序。
- [使用 OpenTelemetry 进行追踪](https://cndoc-langchain.site/langsmith/trace-with-opentelemetry.md): 在 LangSmith 中配置 OpenTelemetry 追踪，包括 LANGSMITH_OTEL_ENABLED 和使用 OpenTelemetry Collector 进行 OTEL 扇出。
- [追踪 Pipecat 应用](https://cndoc-langchain.site/langsmith/trace-with-pipecat.md)
- [追踪 PydanticAI 应用](https://cndoc-langchain.site/langsmith/trace-with-pydantic-ai.md)
- [追踪 Semantic Kernel 应用程序](https://cndoc-langchain.site/langsmith/trace-with-semantic-kernel.md)
- [追踪 Strands Agents 应用](https://cndoc-langchain.site/langsmith/trace-with-strands-agents.md): 使用 LangSmith 追踪 Strands Agents 应用。
- [使用 Temporal 进行追踪](https://cndoc-langchain.site/langsmith/trace-with-temporal.md): 学习如何在 LangSmith 中使用 OpenTelemetry 追踪 Temporal 工作流和活动。
- [追踪 Vercel AI SDK 应用程序（仅限 JS/TS）](https://cndoc-langchain.site/langsmith/trace-with-vercel-ai-sdk.md)
- [无需设置环境变量进行追踪](https://cndoc-langchain.site/langsmith/trace-without-env-vars.md)
- [如何使用轨迹评估来评估您的智能体](https://cndoc-langchain.site/langsmith/trajectory-evals.md)
- [故障排除](https://cndoc-langchain.site/langsmith/troubleshooting.md)
- [Studio 故障排除](https://cndoc-langchain.site/langsmith/troubleshooting-studio.md)
- [排查变量缓存问题](https://cndoc-langchain.site/langsmith/troubleshooting-variable-caching.md)
- [如何使用 REST API 上传在 LangSmith 外部运行的实验](https://cndoc-langchain.site/langsmith/upload-existing-experiments.md)
- [上传带有追踪的文件](https://cndoc-langchain.site/langsmith/upload-files-with-traces.md)
- [如何使用 RemoteGraph 与部署进行交互](https://cndoc-langchain.site/langsmith/use-remote-graph.md)
- [如何使用 Studio](https://cndoc-langchain.site/langsmith/use-studio.md)
- [使用线程](https://cndoc-langchain.site/langsmith/use-threads.md)
- [在提示中使用工具](https://cndoc-langchain.site/langsmith/use-tools.md)
- [使用 Web 钩子](https://cndoc-langchain.site/langsmith/use-webhooks.md)
- [用户管理](https://cndoc-langchain.site/langsmith/user-management.md)
- [使用情况视图](https://cndoc-langchain.site/langsmith/view-usage.md): LangSmith 中提供了哪些使用情况数据、每个指标的含义，以及自托管部署的不同之处。
- [如何使用 Vitest/Jest 运行评估](https://cndoc-langchain.site/langsmith/vitest-jest.md)
- [为规则配置 webhook 通知](https://cndoc-langchain.site/langsmith/webhooks.md): 配置 webhook 通知，以便在自动化规则匹配到 LangSmith 中的新运行时接收 POST 请求。
- [工作负载隔离](https://cndoc-langchain.site/langsmith/workload-isolation.md)
- [使用AI编写提示词](https://cndoc-langchain.site/langsmith/write-prompt-with-ai.md)
- [错误](https://cndoc-langchain.site/oss/javascript/common-errors.md)
- [上下文概述](https://cndoc-langchain.site/oss/javascript/concepts/context.md)
- [记忆概述](https://cndoc-langchain.site/oss/javascript/concepts/memory.md)
- [框架、运行时与工具集](https://cndoc-langchain.site/oss/javascript/concepts/products.md): 了解 LangChain、LangGraph 和 Deep Agents 之间的区别，以及何时使用每种工具
- [提供商与模型](https://cndoc-langchain.site/oss/javascript/concepts/providers-and-models.md): 了解 LangChain 如何使用提供商为您提供来自任何提供商的任何模型的统一 API
- [为代码做贡献](https://cndoc-langchain.site/oss/javascript/contributing/code.md)
- [联合营销](https://cndoc-langchain.site/oss/javascript/contributing/comarketing.md)
- [贡献文档](https://cndoc-langchain.site/oss/javascript/contributing/documentation.md)
- [实现 LangChain 集成](https://cndoc-langchain.site/oss/javascript/contributing/implement-langchain.md)
- [贡献集成](https://cndoc-langchain.site/oss/javascript/contributing/integrations-langchain.md)
- [贡献指南](https://cndoc-langchain.site/oss/javascript/contributing/overview.md)
- [发布集成](https://cndoc-langchain.site/oss/javascript/contributing/publish-langchain.md)
- [使用标准测试](https://cndoc-langchain.site/oss/javascript/contributing/standard-tests-langchain.md)
- [A2A服务器](https://cndoc-langchain.site/oss/javascript/deepagents/a2a.md)
- [代理客户端协议 (ACP)](https://cndoc-langchain.site/oss/javascript/deepagents/acp.md): 通过代理客户端协议 (ACP) 暴露深度代理，以与代码编辑器和 IDE 集成。
- [异步子代理](https://cndoc-langchain.site/oss/javascript/deepagents/async-subagents.md): 启动后台子代理，使其并发运行，同时主管代理继续与用户交互
- [后端](https://cndoc-langchain.site/oss/javascript/deepagents/backends.md): 为 Deep Agents 选择和配置文件系统后端。您可以指定到不同后端的路由、实现虚拟文件系统并强制执行策略。
- [更新日志](https://cndoc-langchain.site/oss/javascript/deepagents/changelog-js.md)
- [配置](https://cndoc-langchain.site/oss/javascript/deepagents/cli/configuration.md): 使用 config.toml、钩子和 MCP 服务器配置 Deep Agents CLI
- [MCP 工具](https://cndoc-langchain.site/oss/javascript/deepagents/cli/mcp-tools.md): 从 MCP（模型上下文协议）服务器加载额外工具
- [Deep Agents CLI](https://cndoc-langchain.site/oss/javascript/deepagents/cli/overview.md): 基于 Deep Agents SDK 构建的终端编码代理
- [模型提供商](https://cndoc-langchain.site/oss/javascript/deepagents/cli/providers.md): 为 Deep Agents CLI 配置任何兼容 LangChain 的模型提供商
- [与 Claude Agent SDK 的比较](https://cndoc-langchain.site/oss/javascript/deepagents/comparison.md): 比较 LangChain Deep Agents 与 Claude Agent SDK，为您的用例选择合适的工具。
- [构建内容生成代理](https://cndoc-langchain.site/oss/javascript/deepagents/content-builder.md): 构建一个具有品牌记忆、技能、子代理和图像生成功能的内容写作代理
- [深度代理中的上下文工程](https://cndoc-langchain.site/oss/javascript/deepagents/context-engineering.md): 控制您的深度代理可以访问哪些上下文，以及如何在长时间运行的任务中管理这些上下文
- [自定义深度代理](https://cndoc-langchain.site/oss/javascript/deepagents/customization.md): 了解如何使用系统提示、工具、子代理等自定义深度代理
- [数据位置](https://cndoc-langchain.site/oss/javascript/deepagents/data-locations.md): Deep Agents CLI 存储配置、会话和自定义文件的位置
- [构建深度研究代理](https://cndoc-langchain.site/oss/javascript/deepagents/deep-research.md): 构建一个具有子代理委派功能的多步骤网络研究代理
- [概览](https://cndoc-langchain.site/oss/javascript/deepagents/frontend/overview.md): 构建显示实时子代理流、任务进度和深度代理沙盒的UI
- [沙箱](https://cndoc-langchain.site/oss/javascript/deepagents/frontend/sandbox.md): 为由沙箱环境支持的编码代理构建类似 IDE 的用户界面
- [子代理流式处理](https://cndoc-langchain.site/oss/javascript/deepagents/frontend/subagent-streaming.md): 显示具有流式内容、进度跟踪和可折叠卡片的专家子代理
- [待办事项列表](https://cndoc-langchain.site/oss/javascript/deepagents/frontend/todo-list.md): 通过从代理状态实时同步的待办事项列表跟踪代理进度
- [投入生产](https://cndoc-langchain.site/oss/javascript/deepagents/going-to-production.md): 通过持久化内存、沙箱、弹性中间件和部署选项，将您的深度代理投入生产
- [Harness 能力](https://cndoc-langchain.site/oss/javascript/deepagents/harness.md)
- [人在回路](https://cndoc-langchain.site/oss/javascript/deepagents/human-in-the-loop.md): 了解如何为敏感工具操作配置人工审批
- [模型上下文协议](https://cndoc-langchain.site/oss/javascript/deepagents/mcp.md)
- [记忆](https://cndoc-langchain.site/oss/javascript/deepagents/memory.md): 为使用 Deep Agents 构建的智能体添加持久记忆，使其能够跨对话学习和改进
- [模型](https://cndoc-langchain.site/oss/javascript/deepagents/models.md): 为 Deep Agents 配置模型提供商和参数
- [Deep Agents 概述](https://cndoc-langchain.site/oss/javascript/deepagents/overview.md): 构建能够规划、使用子代理并利用文件系统处理复杂任务的代理
- [配置文件](https://cndoc-langchain.site/oss/javascript/deepagents/profiles.md): Deep Agents 在选择模型时应用的按提供商和按模型的默认配置包
- [快速开始](https://cndoc-langchain.site/oss/javascript/deepagents/quickstart.md): 在几分钟内构建你的第一个深度代理
- [沙盒](https://cndoc-langchain.site/oss/javascript/deepagents/sandboxes.md): 在隔离环境中使用沙盒后端执行代码
- [技能](https://cndoc-langchain.site/oss/javascript/deepagents/skills.md): 学习如何通过技能扩展深度代理的能力
- [流式处理](https://cndoc-langchain.site/oss/javascript/deepagents/streaming.md): 从深度代理运行和子代理执行中流式传输实时更新
- [子代理](https://cndoc-langchain.site/oss/javascript/deepagents/subagents.md): 了解如何使用子代理来委派工作并保持上下文整洁
- [回调集成](https://cndoc-langchain.site/oss/javascript/integrations/callbacks/index.md): 使用 LangChain JavaScript 进行回调集成。
- [ChatAnthropic 集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/anthropic.md): 使用 LangChain JavaScript 与 ChatAnthropic 聊天模型集成。
- [AzureChatOpenAI 集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/azure.md): 使用 LangChain JavaScript 集成 AzureChatOpenAI 聊天模型。
- [BedrockChat 集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/bedrock.md): 使用 LangChain JavaScript 集成 BedrockChat 聊天模型。
- [ChatGoogle 集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/google.md): 使用 LangChain JavaScript 集成 ChatGoogle 聊天模型。
- [ChatGoogleGenerativeAI 集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/google_generative_ai.md): 使用 LangChain JavaScript 集成 ChatGoogleGenerativeAI 聊天模型。
- [ChatVertexAI 集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/google_vertex_ai.md): 使用 LangChain JavaScript 集成 ChatVertexAI 聊天模型。
- [聊天模型集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/index.md): 使用 LangChain JavaScript 集成聊天模型。
- [ChatOpenAI 集成](https://cndoc-langchain.site/oss/javascript/integrations/chat/openai.md): 使用 LangChain JavaScript 集成 ChatOpenAI 聊天模型。
- [文档加载器集成](https://cndoc-langchain.site/oss/javascript/integrations/document_loaders/index.md): 使用 LangChain JavaScript 集成文档加载器。
- [文档转换器集成](https://cndoc-langchain.site/oss/javascript/integrations/document_transformers/index.md): 使用 LangChain JavaScript 与文档转换器集成。
- [AzureOpenAIEmbeddings 集成](https://cndoc-langchain.site/oss/javascript/integrations/embeddings/azure_openai.md): 使用 LangChain JavaScript 集成 AzureOpenAIEmbeddings 嵌入模型。
- [BedrockEmbeddings 集成](https://cndoc-langchain.site/oss/javascript/integrations/embeddings/bedrock.md): 使用 LangChain JavaScript 集成 BedrockEmbeddings 嵌入模型。
- [嵌入模型集成](https://cndoc-langchain.site/oss/javascript/integrations/embeddings/index.md): 使用 LangChain JavaScript 集成嵌入模型。
- [OpenAIEmbeddings 集成](https://cndoc-langchain.site/oss/javascript/integrations/embeddings/openai.md): 使用 LangChain JavaScript 集成 OpenAIEmbeddings 嵌入模型。
- [缓存集成](https://cndoc-langchain.site/oss/javascript/integrations/llm_caching/index.md): 使用 LangChain JavaScript 集成缓存。
- [大语言模型集成](https://cndoc-langchain.site/oss/javascript/integrations/llms/index.md): 使用 LangChain JavaScript 集成大语言模型。
- [中间件集成](https://cndoc-langchain.site/oss/javascript/integrations/middleware/index.md): 使用 LangChain JavaScript 与中间件集成。
- [所有 LangChain JavaScript 集成提供商](https://cndoc-langchain.site/oss/javascript/integrations/providers/all_providers.md)
- [Anthropic 集成](https://cndoc-langchain.site/oss/javascript/integrations/providers/anthropic.md): 使用 LangChain JavaScript 与 Anthropic 集成。
- [AWS 集成](https://cndoc-langchain.site/oss/javascript/integrations/providers/aws.md): 使用 LangChain JavaScript 与 AWS 集成。
- [Google 集成](https://cndoc-langchain.site/oss/javascript/integrations/providers/google.md): 使用 LangChain JavaScript 与 Google 进行集成。
- [Microsoft 集成](https://cndoc-langchain.site/oss/javascript/integrations/providers/microsoft.md): 使用 LangChain JavaScript 与 Microsoft 进行集成。
- [OpenAI 集成](https://cndoc-langchain.site/oss/javascript/integrations/providers/openai.md): 使用 LangChain JavaScript 与 OpenAI 集成。
- [LangChain JavaScript 集成](https://cndoc-langchain.site/oss/javascript/integrations/providers/overview.md): 使用 LangChain JavaScript/TypeScript 与提供商集成。
- [检索器集成](https://cndoc-langchain.site/oss/javascript/integrations/retrievers/index.md): 使用 LangChain JavaScript 与检索器集成。
- [沙箱集成](https://cndoc-langchain.site/oss/javascript/integrations/sandboxes/index.md): 使用 LangChain JavaScript 与沙箱提供商集成。
- [文本分割器集成](https://cndoc-langchain.site/oss/javascript/integrations/splitters/index.md): 使用 LangChain 集成文本分割器。
- [存储集成](https://cndoc-langchain.site/oss/javascript/integrations/stores/index.md): 使用 LangChain JavaScript 与存储进行集成。
- [Anthropic 集成](https://cndoc-langchain.site/oss/javascript/integrations/tools/anthropic.md): 使用 LangChain JavaScript 与 Anthropic 工具集成。
- [Google 集成](https://cndoc-langchain.site/oss/javascript/integrations/tools/google.md): 使用 LangChain JavaScript 集成 Google Gemini 工具。
- [工具集成](https://cndoc-langchain.site/oss/javascript/integrations/tools/index.md): 使用 LangChain JavaScript 集成工具。
- [OpenAI 集成](https://cndoc-langchain.site/oss/javascript/integrations/tools/openai.md): 使用 LangChain JavaScript 与 OpenAI 工具集成。
- [向量存储集成](https://cndoc-langchain.site/oss/javascript/integrations/vectorstores/index.md): 使用 LangChain JavaScript 集成向量存储。
- [LangChain学院](https://cndoc-langchain.site/oss/javascript/langchain/academy.md)
- [代理](https://cndoc-langchain.site/oss/javascript/langchain/agents.md)
- [更新日志](https://cndoc-langchain.site/oss/javascript/langchain/changelog-js.md)
- [组件架构](https://cndoc-langchain.site/oss/javascript/langchain/component-architecture.md)
- [智能体中的上下文工程](https://cndoc-langchain.site/oss/javascript/langchain/context-engineering.md)
- [LangSmith 部署](https://cndoc-langchain.site/oss/javascript/langchain/deploy.md)
- [分支聊天](https://cndoc-langchain.site/oss/javascript/langchain/frontend/branching-chat.md): 编辑消息、重新生成响应并导航对话分支
- [生成式UI](https://cndoc-langchain.site/oss/javascript/langchain/frontend/generative-ui.md): 使用 json-render 渲染AI生成的用户界面
- [人类在环](https://cndoc-langchain.site/oss/javascript/langchain/frontend/human-in-the-loop.md): 添加基于中断的人工审核审批工作流
- [AI 元素](https://cndoc-langchain.site/oss/javascript/langchain/frontend/integrations/ai-elements.md): 基于 shadcn/ui 的可组合组件，用于 AI 聊天界面，支持 useStream
- [assistant-ui](https://cndoc-langchain.site/oss/javascript/langchain/frontend/integrations/assistant-ui.md): 无头 React AI 聊天框架，具备完整运行时层，通过 useStream 桥接
- [CopilotKit](https://cndoc-langchain.site/oss/javascript/langchain/frontend/integrations/copilotkit.md): 向 LangGraph 部署添加自定义 CopilotKit 端点，并在 React 中渲染结构化生成式 UI
- [OpenUI](https://cndoc-langchain.site/oss/javascript/langchain/frontend/integrations/openui.md): 使用 OpenUI 组件库和 openui-lang 生成完整的交互式仪表板和报告
- [概述](https://cndoc-langchain.site/oss/javascript/langchain/frontend/integrations/overview.md): 将 useStream 连接到任何 React UI 组件库或生成式 UI 框架
- [加入与重连流](https://cndoc-langchain.site/oss/javascript/langchain/frontend/join-rejoin.md): 从运行中的代理流断开连接并重新连接
- [Markdown 消息](https://cndoc-langchain.site/oss/javascript/langchain/frontend/markdown-messages.md): 将 LLM 响应渲染为富文本、格式化的 Markdown，并提供适当的流式支持
- [消息队列](https://cndoc-langchain.site/oss/javascript/langchain/frontend/message-queues.md): 排队多条消息并在代理顺序处理时进行管理
- [概览](https://cndoc-langchain.site/oss/javascript/langchain/frontend/overview.md): 使用来自 LangChain 代理的实时流式传输构建生成式 UI
- [推理令牌](https://cndoc-langchain.site/oss/javascript/langchain/frontend/reasoning-tokens.md): 在可折叠块中展示模型的思考与推理过程
- [结构化输出](https://cndoc-langchain.site/oss/javascript/langchain/frontend/structured-output.md): 使用自定义UI组件渲染结构化的代理响应，而非纯文本
- [时间旅行](https://cndoc-langchain.site/oss/javascript/langchain/frontend/time-travel.md): 检查、导航并从对话历史中的任意检查点恢复执行
- [工具调用](https://cndoc-langchain.site/oss/javascript/langchain/frontend/tool-calling.md): 使用丰富、类型安全的UI卡片显示代理工具调用
- [获取帮助](https://cndoc-langchain.site/oss/javascript/langchain/get-help.md)
- [防护栏](https://cndoc-langchain.site/oss/javascript/langchain/guardrails.md): 为您的代理实施安全检查和内容过滤
- [Human in the Loop](https://cndoc-langchain.site/oss/javascript/langchain/human-in-the-loop.md)
- [安装 LangChain](https://cndoc-langchain.site/oss/javascript/langchain/install.md)
- [使用 LangChain 构建语义搜索引擎](https://cndoc-langchain.site/oss/javascript/langchain/knowledge-base.md)
- [长期记忆](https://cndoc-langchain.site/oss/javascript/langchain/long-term-memory.md): 为 LangChain 代理添加长期记忆，以跨对话和会话存储和召回数据
- [模型上下文协议 (MCP)](https://cndoc-langchain.site/oss/javascript/langchain/mcp.md)
- [消息](https://cndoc-langchain.site/oss/javascript/langchain/messages.md)
- [预构建中间件](https://cndoc-langchain.site/oss/javascript/langchain/middleware/built-in.md): 用于常见智能体用例的预构建中间件
- [自定义中间件](https://cndoc-langchain.site/oss/javascript/langchain/middleware/custom.md)
- [概述](https://cndoc-langchain.site/oss/javascript/langchain/middleware/overview.md): 在每个步骤中控制和自定义代理执行
- [模型](https://cndoc-langchain.site/oss/javascript/langchain/models.md)
- [自定义工作流](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/custom-workflow.md)
- [交接](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/handoffs.md)
- [通过交接构建客户支持](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/handoffs-customer-support.md)
- [多智能体](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/index.md)
- [路由器](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/router.md)
- [使用路由构建多源知识库](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/router-knowledge-base.md)
- [技能](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/skills.md)
- [构建具备按需技能的SQL助手](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/skills-sql-assistant.md)
- [子代理](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/subagents.md)
- [使用子智能体构建个人助手](https://cndoc-langchain.site/oss/javascript/langchain/multi-agent/subagents-personal-assistant.md)
- [LangSmith 可观测性](https://cndoc-langchain.site/oss/javascript/langchain/observability.md)
- [LangChain 概述](https://cndoc-langchain.site/oss/javascript/langchain/overview.md): LangChain 是一个开源框架，提供预构建的智能体架构以及与任何模型或工具的集成——因此您可以构建能够随着生态系统演进快速适应的智能体
- [哲学](https://cndoc-langchain.site/oss/javascript/langchain/philosophy.md): LangChain 的存在是为了让使用大语言模型进行构建变得尽可能简单，同时保持灵活性和生产就绪性。
- [快速入门](https://cndoc-langchain.site/oss/javascript/langchain/quickstart.md): 在几分钟内构建你的第一个智能体
- [使用 LangChain 构建 RAG 代理](https://cndoc-langchain.site/oss/javascript/langchain/rag.md)
- [检索](https://cndoc-langchain.site/oss/javascript/langchain/retrieval.md)
- [运行时](https://cndoc-langchain.site/oss/javascript/langchain/runtime.md)
- [短期记忆](https://cndoc-langchain.site/oss/javascript/langchain/short-term-memory.md)
- [构建 SQL 代理](https://cndoc-langchain.site/oss/javascript/langchain/sql-agent.md)
- [流式处理](https://cndoc-langchain.site/oss/javascript/langchain/streaming.md): 从智能体运行中流式传输实时更新
- [结构化输出](https://cndoc-langchain.site/oss/javascript/langchain/structured-output.md)
- [LangSmith Studio](https://cndoc-langchain.site/oss/javascript/langchain/studio.md)
- [智能体评估](https://cndoc-langchain.site/oss/javascript/langchain/test/evals.md): 使用 AgentEvals 和 LangSmith，通过确定性匹配或 LLM 作为评判的评估器来评估智能体轨迹。
- [测试](https://cndoc-langchain.site/oss/javascript/langchain/test/index.md): 测试 LangChain 智能体的策略，包括单元测试、集成测试和轨迹评估。
- [集成测试](https://cndoc-langchain.site/oss/javascript/langchain/test/integration-testing.md): 通过组织测试、管理密钥、处理不稳定性和控制成本，使用真实LLM API测试代理。
- [单元测试](https://cndoc-langchain.site/oss/javascript/langchain/test/unit-testing.md): 使用假聊天模型和内存持久化测试智能体逻辑，无需调用 API。
- [工具](https://cndoc-langchain.site/oss/javascript/langchain/tools.md)
- [代理聊天界面](https://cndoc-langchain.site/oss/javascript/langchain/ui.md)
- [使用 LangChain 构建语音代理](https://cndoc-langchain.site/oss/javascript/langchain/voice-agent.md)
- [内存](https://cndoc-langchain.site/oss/javascript/langgraph/add-memory.md)
- [使用 LangGraph 构建自定义 RAG 代理](https://cndoc-langchain.site/oss/javascript/langgraph/agentic-rag.md)
- [应用结构](https://cndoc-langchain.site/oss/javascript/langgraph/application-structure.md)
- [向后兼容性](https://cndoc-langchain.site/oss/javascript/langgraph/backward-compatibility.md): 在生产环境中更新 LangGraph 图代码，同时不中断正在进行的运行。
- [案例研究](https://cndoc-langchain.site/oss/javascript/langgraph/case-studies.md)
- [更新日志](https://cndoc-langchain.site/oss/javascript/langgraph/changelog-js.md)
- [在 Graph API 和 Functional API 之间选择](https://cndoc-langchain.site/oss/javascript/langgraph/choosing-apis.md)
- [LangSmith 部署](https://cndoc-langchain.site/oss/javascript/langgraph/deploy.md)
- [持久化执行](https://cndoc-langchain.site/oss/javascript/langgraph/durable-execution.md)
- [图执行](https://cndoc-langchain.site/oss/javascript/langgraph/frontend/graph-execution.md): 可视化多步骤图流水线，展示每个节点的状态和流式内容
- [概述](https://cndoc-langchain.site/oss/javascript/langgraph/frontend/overview.md): 将LangGraph代理渲染到前端
- [功能 API 概览](https://cndoc-langchain.site/oss/javascript/langgraph/functional-api.md)
- [图 API 概览](https://cndoc-langchain.site/oss/javascript/langgraph/graph-api.md)
- [安装 LangGraph](https://cndoc-langchain.site/oss/javascript/langgraph/install.md)
- [中断](https://cndoc-langchain.site/oss/javascript/langgraph/interrupts.md)
- [运行本地服务器](https://cndoc-langchain.site/oss/javascript/langgraph/local-server.md)
- [LangSmith 可观测性](https://cndoc-langchain.site/oss/javascript/langgraph/observability.md)
- [LangGraph 概述](https://cndoc-langchain.site/oss/javascript/langgraph/overview.md): 使用 LangGraph 掌控能力，设计能够可靠处理复杂任务的智能体
- [持久化](https://cndoc-langchain.site/oss/javascript/langgraph/persistence.md)
- [LangGraph 运行时](https://cndoc-langchain.site/oss/javascript/langgraph/pregel.md)
- [快速入门](https://cndoc-langchain.site/oss/javascript/langgraph/quickstart.md)
- [流式处理](https://cndoc-langchain.site/oss/javascript/langgraph/streaming.md)
- [LangSmith Studio](https://cndoc-langchain.site/oss/javascript/langgraph/studio.md)
- [测试](https://cndoc-langchain.site/oss/javascript/langgraph/test.md)
- [用 LangGraph 思考](https://cndoc-langchain.site/oss/javascript/langgraph/thinking-in-langgraph.md): 学习如何思考使用 LangGraph 构建代理
- [代理聊天界面](https://cndoc-langchain.site/oss/javascript/langgraph/ui.md)
- [使用函数式 API](https://cndoc-langchain.site/oss/javascript/langgraph/use-functional-api.md)
- [使用图 API](https://cndoc-langchain.site/oss/javascript/langgraph/use-graph-api.md)
- [子图](https://cndoc-langchain.site/oss/javascript/langgraph/use-subgraphs.md)
- [使用时间旅行](https://cndoc-langchain.site/oss/javascript/langgraph/use-time-travel.md): 在LangGraph中重放过去的执行并分叉以探索替代路径
- [工作流与代理](https://cndoc-langchain.site/oss/javascript/langgraph/workflows-agents.md)
- [学习](https://cndoc-langchain.site/oss/javascript/learn.md): 教程、概念指南和资源，助您快速入门。
- [LangChain v1 迁移指南](https://cndoc-langchain.site/oss/javascript/migrate/langchain-v1.md)
- [LangGraph v1 迁移指南](https://cndoc-langchain.site/oss/javascript/migrate/langgraph-v1.md)
- [深度代理](https://cndoc-langchain.site/oss/javascript/reference/deepagents-javascript.md)
- [集成](https://cndoc-langchain.site/oss/javascript/reference/integrations-javascript.md)
- [LangChain SDK](https://cndoc-langchain.site/oss/javascript/reference/langchain-javascript.md)
- [LangGraph SDK](https://cndoc-langchain.site/oss/javascript/reference/langgraph-javascript.md)
- [参考文档](https://cndoc-langchain.site/oss/javascript/reference/overview.md)
- [发布策略](https://cndoc-langchain.site/oss/javascript/release-policy.md)
- [更新日志](https://cndoc-langchain.site/oss/javascript/releases/changelog.md): 我们 JavaScript/TypeScript 包的更新与改进日志
- [LangChain v1 新特性](https://cndoc-langchain.site/oss/javascript/releases/langchain-v1.md)
- [LangGraph v1 新特性](https://cndoc-langchain.site/oss/javascript/releases/langgraph-v1.md)
- [安全策略](https://cndoc-langchain.site/oss/javascript/security-policy.md)
- [版本控制](https://cndoc-langchain.site/oss/javascript/versioning.md)
- [错误](https://cndoc-langchain.site/oss/python/common-errors.md)
- [上下文概述](https://cndoc-langchain.site/oss/python/concepts/context.md)
- [记忆概述](https://cndoc-langchain.site/oss/python/concepts/memory.md)
- [框架、运行时与工具集](https://cndoc-langchain.site/oss/python/concepts/products.md): 了解 LangChain、LangGraph 和 Deep Agents 之间的区别以及何时使用它们
- [提供商与模型](https://cndoc-langchain.site/oss/python/concepts/providers-and-models.md): 了解 LangChain 如何使用提供商为您提供一个统一的 API，以访问来自任何提供商的模型
- [代码贡献](https://cndoc-langchain.site/oss/python/contributing/code.md)
- [联合营销](https://cndoc-langchain.site/oss/python/contributing/comarketing.md)
- [贡献文档](https://cndoc-langchain.site/oss/python/contributing/documentation.md)
- [实现 LangChain 集成](https://cndoc-langchain.site/oss/python/contributing/implement-langchain.md)
- [贡献集成](https://cndoc-langchain.site/oss/python/contributing/integrations-langchain.md)
- [贡献指南](https://cndoc-langchain.site/oss/python/contributing/overview.md)
- [发布集成](https://cndoc-langchain.site/oss/python/contributing/publish-langchain.md)
- [使用标准测试](https://cndoc-langchain.site/oss/python/contributing/standard-tests-langchain.md)
- [A2A服务器](https://cndoc-langchain.site/oss/python/deepagents/a2a.md)
- [代理客户端协议 (ACP)](https://cndoc-langchain.site/oss/python/deepagents/acp.md): 通过代理客户端协议 (ACP) 暴露深度代理，以与代码编辑器和 IDE 集成。
- [异步子代理](https://cndoc-langchain.site/oss/python/deepagents/async-subagents.md): 启动在后台并发运行的子代理，同时主管代理继续与用户交互
- [后端](https://cndoc-langchain.site/oss/python/deepagents/backends.md): 为 Deep Agents 选择和配置文件系统后端。您可以指定到不同后端的路由、实现虚拟文件系统并强制执行策略。
- [更新日志](https://cndoc-langchain.site/oss/python/deepagents/changelog-py.md)
- [配置](https://cndoc-langchain.site/oss/python/deepagents/cli/configuration.md): 使用 config.toml、钩子和 MCP 服务器配置 Deep Agents CLI
- [MCP 工具](https://cndoc-langchain.site/oss/python/deepagents/cli/mcp-tools.md): 从 MCP（模型上下文协议）服务器加载额外工具
- [Deep Agents CLI](https://cndoc-langchain.site/oss/python/deepagents/cli/overview.md): 基于 Deep Agents SDK 构建的终端编码代理
- [模型提供商](https://cndoc-langchain.site/oss/python/deepagents/cli/providers.md): 为 Deep Agents CLI 配置任何兼容 LangChain 的模型提供商
- [与 Claude Agent SDK 的比较](https://cndoc-langchain.site/oss/python/deepagents/comparison.md): 比较 LangChain Deep Agents 与 Claude Agent SDK，为您的用例选择合适的工具。
- [构建内容构建代理](https://cndoc-langchain.site/oss/python/deepagents/content-builder.md): 构建一个具有品牌记忆、技能、子代理和图像生成功能的内容写作代理
- [深度代理中的上下文工程](https://cndoc-langchain.site/oss/python/deepagents/context-engineering.md): 控制您的深度代理可以访问哪些上下文，以及如何在长时间运行的任务中管理这些上下文
- [自定义 Deep Agents](https://cndoc-langchain.site/oss/python/deepagents/customization.md): 了解如何使用系统提示、工具、子代理等自定义 Deep Agents
- [构建数据分析代理](https://cndoc-langchain.site/oss/python/deepagents/data-analysis.md): 构建一个分析数据文件、生成可视化并共享结果的代理
- [数据位置](https://cndoc-langchain.site/oss/python/deepagents/data-locations.md): Deep Agents CLI 存储配置、会话和自定义文件的位置
- [构建深度研究代理](https://cndoc-langchain.site/oss/python/deepagents/deep-research.md): 构建一个具有子代理委派功能的多步骤网络研究代理
- [使用 CLI 部署](https://cndoc-langchain.site/oss/python/deepagents/deploy.md): 使用 Deep Agents CLI 将模型无关的开源智能体部署到生产环境
- [概览](https://cndoc-langchain.site/oss/python/deepagents/frontend/overview.md): 构建显示实时子代理流、任务进度和沙盒的 Deep Agents 用户界面
- [沙箱](https://cndoc-langchain.site/oss/python/deepagents/frontend/sandbox.md): 为编码代理构建一个类似 IDE 的界面，由沙箱环境支持
- [子代理流式处理](https://cndoc-langchain.site/oss/python/deepagents/frontend/subagent-streaming.md): 显示具有流式内容、进度跟踪和可折叠卡片的专家子代理
- [待办事项列表](https://cndoc-langchain.site/oss/python/deepagents/frontend/todo-list.md): 通过从智能体状态同步的实时待办事项列表跟踪智能体进度
- [投入生产](https://cndoc-langchain.site/oss/python/deepagents/going-to-production.md): 通过持久化内存、沙箱、弹性中间件和部署选项，将您的深度智能体投入生产
- [框架能力](https://cndoc-langchain.site/oss/python/deepagents/harness.md)
- [人机交互](https://cndoc-langchain.site/oss/python/deepagents/human-in-the-loop.md): 了解如何为敏感工具操作配置人工审批
- [模型上下文协议](https://cndoc-langchain.site/oss/python/deepagents/mcp.md)
- [记忆](https://cndoc-langchain.site/oss/python/deepagents/memory.md): 为使用 Deep Agents 构建的代理添加持久记忆，使其能够跨对话学习和改进
- [模型](https://cndoc-langchain.site/oss/python/deepagents/models.md): 为 Deep Agents 配置模型提供商和参数
- [Deep Agents 概述](https://cndoc-langchain.site/oss/python/deepagents/overview.md): 构建能够规划、使用子代理并利用文件系统处理复杂任务的代理
- [权限](https://cndoc-langchain.site/oss/python/deepagents/permissions.md): 使用声明式权限规则控制 Deep Agents 的文件系统访问
- [配置文件](https://cndoc-langchain.site/oss/python/deepagents/profiles.md): Deep Agents 在选择模型时应用的按提供商和按模型打包的默认设置
- [快速入门](https://cndoc-langchain.site/oss/python/deepagents/quickstart.md): 在几分钟内构建你的第一个深度代理
- [沙盒](https://cndoc-langchain.site/oss/python/deepagents/sandboxes.md): 使用沙盒后端在隔离环境中执行代码
- [技能](https://cndoc-langchain.site/oss/python/deepagents/skills.md): 了解如何通过技能扩展深度代理的能力
- [流式处理](https://cndoc-langchain.site/oss/python/deepagents/streaming.md): 从深度代理运行和子代理执行中流式传输实时更新
- [子代理](https://cndoc-langchain.site/oss/python/deepagents/subagents.md): 了解如何使用子代理来委派工作并保持上下文整洁
- [聊天模型集成](https://cndoc-langchain.site/oss/python/integrations/chat/index.md): 使用 LangChain Python 集成聊天模型。
- [检查点器集成](https://cndoc-langchain.site/oss/python/integrations/checkpointers/index.md): 与 LangGraph 持久化集成的检查点器后端。
- [文档加载器集成](https://cndoc-langchain.site/oss/python/integrations/document_loaders/index.md): 使用 LangChain Python 集成文档加载器。
- [嵌入模型集成](https://cndoc-langchain.site/oss/python/integrations/embeddings/index.md): 使用 LangChain Python 集成嵌入模型。
- [中间件集成](https://cndoc-langchain.site/oss/python/integrations/middleware/index.md): 使用 LangChain Python 与中间件集成。
- [所有 LangChain Python 集成提供商](https://cndoc-langchain.site/oss/python/integrations/providers/all_providers.md)
- [Anthropic (Claude) 集成](https://cndoc-langchain.site/oss/python/integrations/providers/anthropic.md): 使用 LangChain Python 与 Anthropic (Claude) 进行集成。
- [AWS（亚马逊）集成](https://cndoc-langchain.site/oss/python/integrations/providers/aws.md): 使用 LangChain Python 与 AWS（亚马逊）集成。
- [Google 集成](https://cndoc-langchain.site/oss/python/integrations/providers/google.md): 使用 LangChain Python 与 Google 集成。
- [Groq 集成](https://cndoc-langchain.site/oss/python/integrations/providers/groq.md): 使用 LangChain Python 与 Groq 集成。
- [Hugging Face 集成](https://cndoc-langchain.site/oss/python/integrations/providers/huggingface.md): 使用 LangChain Python 与 Hugging Face 集成。
- [Microsoft 集成](https://cndoc-langchain.site/oss/python/integrations/providers/microsoft.md): 使用 LangChain Python 与 Microsoft 集成。
- [NVIDIA](https://cndoc-langchain.site/oss/python/integrations/providers/nvidia.md): 使用 LangChain Python 与 NVIDIA 集成。
- [Ollama 集成](https://cndoc-langchain.site/oss/python/integrations/providers/ollama.md): 使用 LangChain Python 与 Ollama 集成。
- [OpenAI 集成](https://cndoc-langchain.site/oss/python/integrations/providers/openai.md): 使用 LangChain Python 与 OpenAI 集成。
- [LangChain Python 集成](https://cndoc-langchain.site/oss/python/integrations/providers/overview.md): 使用 LangChain Python 与提供商集成。
- [检索器集成](https://cndoc-langchain.site/oss/python/integrations/retrievers/index.md): 使用 LangChain Python 与检索器集成。
- [沙箱集成](https://cndoc-langchain.site/oss/python/integrations/sandboxes/index.md): 使用 LangChain Python 与沙箱提供商集成。
- [文本分割器集成](https://cndoc-langchain.site/oss/python/integrations/splitters/index.md): 使用 LangChain 集成文本分割器。
- [工具集成](https://cndoc-langchain.site/oss/python/integrations/tools/index.md): 使用 LangChain Python 集成工具。
- [向量存储集成](https://cndoc-langchain.site/oss/python/integrations/vectorstores/index.md): 使用 LangChain Python 与向量存储集成。
- [LangChain学院](https://cndoc-langchain.site/oss/python/langchain/academy.md)
- [代理](https://cndoc-langchain.site/oss/python/langchain/agents.md)
- [更新日志](https://cndoc-langchain.site/oss/python/langchain/changelog-py.md)
- [组件架构](https://cndoc-langchain.site/oss/python/langchain/component-architecture.md)
- [智能体中的上下文工程](https://cndoc-langchain.site/oss/python/langchain/context-engineering.md)
- [LangSmith 部署](https://cndoc-langchain.site/oss/python/langchain/deploy.md)
- [分支聊天](https://cndoc-langchain.site/oss/python/langchain/frontend/branching-chat.md): 编辑消息、重新生成响应并导航对话分支
- [生成式用户界面](https://cndoc-langchain.site/oss/python/langchain/frontend/generative-ui.md): 使用 json-render 渲染 AI 生成的用户界面
- [人类在环](https://cndoc-langchain.site/oss/python/langchain/frontend/human-in-the-loop.md): 添加基于中断的人工审核审批工作流
- [AI 元素](https://cndoc-langchain.site/oss/python/langchain/frontend/integrations/ai-elements.md): 基于 shadcn/ui 的可组合组件，用于构建支持 useStream 的 AI 聊天界面
- [assistant-ui](https://cndoc-langchain.site/oss/python/langchain/frontend/integrations/assistant-ui.md): 一个无头的 React AI 聊天框架，具备完整的运行时层，通过 useStream 进行桥接
- [CopilotKit](https://cndoc-langchain.site/oss/python/langchain/frontend/integrations/copilotkit.md): 向 LangGraph 部署添加自定义 CopilotKit 端点，并在 React 中渲染结构化生成式 UI
- [OpenUI](https://cndoc-langchain.site/oss/python/langchain/frontend/integrations/openui.md): 使用 OpenUI 组件库和 openui-lang 生成完整的交互式仪表板和报告
- [概述](https://cndoc-langchain.site/oss/python/langchain/frontend/integrations/overview.md): 将 useStream 连接到任何 React UI 组件库或生成式 UI 框架
- [加入与重新加入流](https://cndoc-langchain.site/oss/python/langchain/frontend/join-rejoin.md): 从运行中的代理流断开连接并重新连接
- [Markdown 消息](https://cndoc-langchain.site/oss/python/langchain/frontend/markdown-messages.md): 将 LLM 响应渲染为格式丰富的 Markdown，并提供适当的流式支持
- [消息队列](https://cndoc-langchain.site/oss/python/langchain/frontend/message-queues.md): 排队多条消息并在代理顺序处理时进行管理
- [概览](https://cndoc-langchain.site/oss/python/langchain/frontend/overview.md): 使用来自 LangChain 代理的实时流构建生成式 UI
- [推理令牌](https://cndoc-langchain.site/oss/python/langchain/frontend/reasoning-tokens.md): 在可折叠块中显示模型的思考和推理过程
- [结构化输出](https://cndoc-langchain.site/oss/python/langchain/frontend/structured-output.md): 使用自定义UI组件渲染结构化代理响应，而非纯文本
- [时间旅行](https://cndoc-langchain.site/oss/python/langchain/frontend/time-travel.md): 检查、导航并从对话历史中的任意检查点恢复
- [工具调用](https://cndoc-langchain.site/oss/python/langchain/frontend/tool-calling.md): 使用丰富、类型安全的 UI 卡片显示代理工具调用
- [获取帮助](https://cndoc-langchain.site/oss/python/langchain/get-help.md)
- [防护栏](https://cndoc-langchain.site/oss/python/langchain/guardrails.md): 为您的代理实施安全检查和内容过滤
- [Human in the Loop](https://cndoc-langchain.site/oss/python/langchain/human-in-the-loop.md)
- [安装 LangChain](https://cndoc-langchain.site/oss/python/langchain/install.md)
- [使用 LangChain 构建语义搜索引擎](https://cndoc-langchain.site/oss/python/langchain/knowledge-base.md)
- [长期记忆](https://cndoc-langchain.site/oss/python/langchain/long-term-memory.md): 为 LangChain 代理添加长期记忆，以跨对话和会话存储和召回数据
- [模型上下文协议 (MCP)](https://cndoc-langchain.site/oss/python/langchain/mcp.md)
- [消息](https://cndoc-langchain.site/oss/python/langchain/messages.md)
- [预构建中间件](https://cndoc-langchain.site/oss/python/langchain/middleware/built-in.md): 针对常见智能体用例的预构建中间件
- [自定义中间件](https://cndoc-langchain.site/oss/python/langchain/middleware/custom.md)
- [概览](https://cndoc-langchain.site/oss/python/langchain/middleware/overview.md): 在每一步控制和自定义代理执行
- [模型](https://cndoc-langchain.site/oss/python/langchain/models.md)
- [自定义工作流](https://cndoc-langchain.site/oss/python/langchain/multi-agent/custom-workflow.md)
- [交接](https://cndoc-langchain.site/oss/python/langchain/multi-agent/handoffs.md)
- [通过交接构建客户支持](https://cndoc-langchain.site/oss/python/langchain/multi-agent/handoffs-customer-support.md)
- [多智能体](https://cndoc-langchain.site/oss/python/langchain/multi-agent/index.md)
- [路由器](https://cndoc-langchain.site/oss/python/langchain/multi-agent/router.md)
- [使用路由构建多源知识库](https://cndoc-langchain.site/oss/python/langchain/multi-agent/router-knowledge-base.md)
- [技能](https://cndoc-langchain.site/oss/python/langchain/multi-agent/skills.md)
- [使用按需技能构建 SQL 助手](https://cndoc-langchain.site/oss/python/langchain/multi-agent/skills-sql-assistant.md)
- [子代理](https://cndoc-langchain.site/oss/python/langchain/multi-agent/subagents.md)
- [使用子代理构建个人助手](https://cndoc-langchain.site/oss/python/langchain/multi-agent/subagents-personal-assistant.md)
- [LangSmith 可观测性](https://cndoc-langchain.site/oss/python/langchain/observability.md)
- [LangChain 概述](https://cndoc-langchain.site/oss/python/langchain/overview.md): LangChain 是一个开源框架，具有预构建的代理架构以及与任何模型或工具的集成——因此您可以构建能够随着生态系统演进而快速适应的代理
- [哲学](https://cndoc-langchain.site/oss/python/langchain/philosophy.md): LangChain 的存在是为了让使用大语言模型进行构建变得尽可能简单，同时保持灵活性和生产就绪性。
- [快速入门](https://cndoc-langchain.site/oss/python/langchain/quickstart.md): 在几分钟内构建你的第一个代理
- [使用 LangChain 构建 RAG 代理](https://cndoc-langchain.site/oss/python/langchain/rag.md)
- [检索](https://cndoc-langchain.site/oss/python/langchain/retrieval.md)
- [运行时](https://cndoc-langchain.site/oss/python/langchain/runtime.md)
- [短期记忆](https://cndoc-langchain.site/oss/python/langchain/short-term-memory.md)
- [构建 SQL 代理](https://cndoc-langchain.site/oss/python/langchain/sql-agent.md)
- [流式处理](https://cndoc-langchain.site/oss/python/langchain/streaming.md): 从代理运行中流式传输实时更新
- [结构化输出](https://cndoc-langchain.site/oss/python/langchain/structured-output.md)
- [LangSmith Studio](https://cndoc-langchain.site/oss/python/langchain/studio.md)
- [Agent 评估](https://cndoc-langchain.site/oss/python/langchain/test/evals.md): 使用 AgentEvals 和 LangSmith，通过确定性匹配或 LLM 作为评判的评估器来评估 Agent 轨迹。
- [测试](https://cndoc-langchain.site/oss/python/langchain/test/index.md): 测试 LangChain 代理的策略，包括单元测试、集成测试和轨迹评估。
- [集成测试](https://cndoc-langchain.site/oss/python/langchain/test/integration-testing.md): 通过组织测试、管理密钥、处理不稳定性和控制成本，使用真实 LLM API 测试智能体。
- [单元测试](https://cndoc-langchain.site/oss/python/langchain/test/unit-testing.md): 使用模拟聊天模型和内存持久化，在不调用 API 的情况下测试智能体逻辑。
- [工具](https://cndoc-langchain.site/oss/python/langchain/tools.md)
- [代理聊天界面](https://cndoc-langchain.site/oss/python/langchain/ui.md)
- [使用 LangChain 构建语音代理](https://cndoc-langchain.site/oss/python/langchain/voice-agent.md)
- [内存](https://cndoc-langchain.site/oss/python/langgraph/add-memory.md)
- [使用 LangGraph 构建自定义 RAG 代理](https://cndoc-langchain.site/oss/python/langgraph/agentic-rag.md)
- [应用结构](https://cndoc-langchain.site/oss/python/langgraph/application-structure.md)
- [向后兼容性](https://cndoc-langchain.site/oss/python/langgraph/backward-compatibility.md): 在生产环境中更新 LangGraph 图代码，同时不破坏进行中的运行。
- [案例研究](https://cndoc-langchain.site/oss/python/langgraph/case-studies.md)
- [更新日志](https://cndoc-langchain.site/oss/python/langgraph/changelog-py.md)
- [在图 API 与函数式 API 之间选择](https://cndoc-langchain.site/oss/python/langgraph/choosing-apis.md)
- [LangSmith 部署](https://cndoc-langchain.site/oss/python/langgraph/deploy.md)
- [持久化执行](https://cndoc-langchain.site/oss/python/langgraph/durable-execution.md)
- [容错](https://cndoc-langchain.site/oss/python/langgraph/fault-tolerance.md): 在 LangGraph 中配置节点级超时、重试和错误处理器。
- [图执行](https://cndoc-langchain.site/oss/python/langgraph/frontend/graph-execution.md): 可视化多步骤图流水线，展示每个节点的状态和流式内容
- [概览](https://cndoc-langchain.site/oss/python/langgraph/frontend/overview.md): 将LangGraph代理渲染到前端
- [功能 API 概览](https://cndoc-langchain.site/oss/python/langgraph/functional-api.md)
- [图 API 概览](https://cndoc-langchain.site/oss/python/langgraph/graph-api.md)
- [安装 LangGraph](https://cndoc-langchain.site/oss/python/langgraph/install.md)
- [中断](https://cndoc-langchain.site/oss/python/langgraph/interrupts.md)
- [运行本地服务器](https://cndoc-langchain.site/oss/python/langgraph/local-server.md)
- [LangSmith 可观测性](https://cndoc-langchain.site/oss/python/langgraph/observability.md)
- [LangGraph 概述](https://cndoc-langchain.site/oss/python/langgraph/overview.md): 使用 LangGraph 掌控能力，设计能够可靠处理复杂任务的智能体
- [持久化](https://cndoc-langchain.site/oss/python/langgraph/persistence.md)
- [LangGraph 运行时](https://cndoc-langchain.site/oss/python/langgraph/pregel.md)
- [快速入门](https://cndoc-langchain.site/oss/python/langgraph/quickstart.md)
- [构建自定义 SQL 代理](https://cndoc-langchain.site/oss/python/langgraph/sql-agent.md)
- [流式处理](https://cndoc-langchain.site/oss/python/langgraph/streaming.md)
- [LangSmith Studio](https://cndoc-langchain.site/oss/python/langgraph/studio.md)
- [测试](https://cndoc-langchain.site/oss/python/langgraph/test.md)
- [用LangGraph思考](https://cndoc-langchain.site/oss/python/langgraph/thinking-in-langgraph.md): 学习如何用LangGraph思考构建智能体
- [Agent Chat UI](https://cndoc-langchain.site/oss/python/langgraph/ui.md)
- [使用功能API](https://cndoc-langchain.site/oss/python/langgraph/use-functional-api.md)
- [使用图 API](https://cndoc-langchain.site/oss/python/langgraph/use-graph-api.md)
- [子图](https://cndoc-langchain.site/oss/python/langgraph/use-subgraphs.md)
- [使用时间旅行](https://cndoc-langchain.site/oss/python/langgraph/use-time-travel.md): 在LangGraph中重放过去的执行并分叉以探索替代路径
- [工作流与代理](https://cndoc-langchain.site/oss/python/langgraph/workflows-agents.md)
- [学习](https://cndoc-langchain.site/oss/python/learn.md): 教程、概念指南和资源，助您快速上手。
- [LangChain v1 迁移指南](https://cndoc-langchain.site/oss/python/migrate/langchain-v1.md)
- [LangGraph v1 迁移指南](https://cndoc-langchain.site/oss/python/migrate/langgraph-v1.md)
- [深度代理](https://cndoc-langchain.site/oss/python/reference/deepagents-python.md)
- [集成](https://cndoc-langchain.site/oss/python/reference/integrations-python.md)
- [LangChain SDK](https://cndoc-langchain.site/oss/python/reference/langchain-python.md)
- [LangGraph SDK](https://cndoc-langchain.site/oss/python/reference/langgraph-python.md)
- [参考文档](https://cndoc-langchain.site/oss/python/reference/overview.md)
- [发布策略](https://cndoc-langchain.site/oss/python/release-policy.md)
- [更新日志](https://cndoc-langchain.site/oss/python/releases/changelog.md): 我们 Python 包的更新和改进日志
- [LangChain v1 新特性](https://cndoc-langchain.site/oss/python/releases/langchain-v1.md)
- [LangGraph v1 新特性](https://cndoc-langchain.site/oss/python/releases/langgraph-v1.md)
- [安全策略](https://cndoc-langchain.site/oss/python/security-policy.md)
- [版本控制](https://cndoc-langchain.site/oss/python/versioning.md)

## OpenAPI Specs

- [openapi](https://cndoc-langchain.site/api-reference/openapi.json)
- [agent-server-openapi](https://cndoc-langchain.site/langsmith/agent-server-openapi.json)