Wardn AI是受管AI工具访问的控制平面.

它给团队一个地方批准 模式背景协议服务器,连接 秘密和模型提供者,运行助手, 并审查每一个工具呼叫,批准, 成本和运行时间问题。 代理仍然可以使用 MCP 生态系统,但决定 关于信任、证书、政策和可审计性,应留在沃登而不是 分布于本地配置文件和一次性bots.

产品领域 典型的工作流程 Docker 作曲 本地开发 生产说明

□ 赞助者

已删除 为开发商提供新闻、辅导和发行的个人阅读信息。 跟随你最喜欢的来源和话题在一个地方。 华德AI的骄傲赞助商. 探索DevFeed —

□ 文档

沃德的公共文献来源生活于docs/. 它使用 Antora 和 AsciiDoc 建立一个静态的、版本化的文档网站。

根包还暴露出[[T1],npm run docs:preview,以及 npm run docs:sync:openapi.

∮沃德的本意∮

AI代理在可以跨公司系统行事时变得有用:源控制, 门票,文档,数据存储,客户工具,以及内部API. 难的是

Wardn AI

Wardn AI is a control plane for governed AI tool access.

It gives teams one place to approve Model Context Protocol (MCP) servers, connect secrets and model providers, run assistants, and review every tool call, approval, cost, and runtime issue. Agents can still use the MCP ecosystem, but the decisions about trust, credentials, policy, and auditability stay in Wardn instead of being spread across local config files and one-off bots.

Product Areas | Typical Workflow | Docker Compose | Local Development | Production Notes

Sponsored by

Documentation

Wardn’s public documentation source lives in docs/. It uses Antora and AsciiDoc to build a static, versioned docs site.

cd docs
npm ci
npm run build
npm run preview

The root package also exposes npm run docs:build, npm run docs:preview, and npm run docs:sync:openapi.

What Wardn Is For

AI agents become useful when they can act across company systems: source control, tickets, documents, data stores, customer tools, and internal APIs. The hard part is not only connecting those tools. Teams also need to know which tools are approved, who can use them, which credentials are in play, when a human must approve an action, and what happened after an agent ran.

Wardn is built for that operating layer:

• Curate a trusted catalog of MCP servers for each organization.
• Install and validate tool servers inside scoped workspaces.
• Bind approved tools to agents instead of exposing broad credentials directly.
• Apply access rules, limits, and human approvals before risky work runs.
• Run agents through chat, scheduled tasks, and connected chat channels.
• Trace LLM usage, tool calls, runtime sessions, errors, and approvals from one UI.

Product Areas

Trusted MCP Catalog

Wardn gives platform teams a governed registry for MCP servers and versions. Teams can add, review, edit, and sync server metadata, then install approved servers into the workspaces that need them.

Workspace Agents

Workspace agents use configured LLM providers and the MCP servers assigned to them.