ADOS MCP
Enables AI clients to read live drone status, telemetry, and parameters, and to control drone operations with scoped permissions, confirmation steps, and audit trails.
README
ADOS MCP
A Model Context Protocol server for the ADOS drone platform. Read and control a drone or a whole fleet from your AI client, with a scope on every connection, a confirmation on anything that moves an aircraft, and an audit trail on every call.
Part of the ADOS ecosystem. Pairs with the ADOS Drone Agent, ADOS Mission Control, and ADOS Extensions.
<p align="center"> <a href="https://docs.altnautica.com">Docs</a> · <a href="https://github.com/altnautica/ADOS-MCP/issues">Issues</a> · <a href="https://modelcontextprotocol.io">Model Context Protocol</a> </p>
What it is
ADOS MCP lets any MCP-capable AI client, Claude Code first-class, plus Claude Desktop, Cursor, VS Code, or a plain HTTP client, connect to an ADOS drone or ground station and:
- Read live status, telemetry, and health for a node or a fleet.
- Read every flight-controller parameter with full meaning and get tuning recommendations.
- Administer the platform: rename a node, update the software, install and configure plugins, restart a service, manage pairing and network settings.
- Query logs and a full audit trail.
- Fly, when the operator explicitly grants it: arm, change mode, go to a point, run a mission, land, return home, or stop.
The drone is the server; the AI is the client. No model runs on the drone. The server is a thin layer over interfaces that already exist, so behavior is defined once and stays consistent whether it is reached from the console, the command line, or an AI client.
Visibility is broad; action is gated. Reading is open within the granted scope. Writing is controlled by that scope, a safety class per action, a confirmation step for anything that moves an aircraft or changes state, and an operator-present check for flight. Every call, allowed or denied, is recorded.
Two ways to run it
You run the server yourself, on your own machine, in one of two modes:
| Mode | Command | Reaches |
|---|---|---|
| Fleet-mode | --target fleet --gcs prod |
your Mission Control fleet (its cloud-connected drones) |
| Agent-mode | --target agent <host> |
one drone directly on the LAN |
Both expose the identical tools; only the reach differs.
- Fleet-mode is the primary pathway. The server connects to a Mission Control (GCS) backend — your own local one (
--gcs local) or the production one (--gcs prod) — as you, the operator, and reaches the drones that Mission Control manages. It is the AI-native interface to your fleet. (Reach limit: Mission Control tracks cloud-connected drones; a drone that is only on your LAN is reached in agent-mode.) - Agent-mode is the local-first direct pathway. The server points straight at one drone's agent on the LAN — no cloud round-trip — the shape a bench or field setup uses.
Quick start (Claude Code)
Sign in and mint one scoped machine credential in the Mission Control MCP tab, then run the server on your machine:
# Your whole fleet, through Mission Control (production backend):
export ADOS_MCP_TOKEN="paste-the-machine-credential-from-the-tab"
claude mcp add ados -- npx -y @altnautica/ados-mcp --target fleet --gcs prod
# One drone, directly on the LAN:
claude mcp add ados-lan -- npx -y @altnautica/ados-mcp --target agent <host>
The credential is scoped and revocable, and stored on the backend only as a hash. Revoke it in the tab to cut a server off instantly.
Then ask your client for fleet status. See docs/ for the full connect recipes and the tool catalog.
Transports
- Streamable HTTP (
POST /mcp, single endpoint, SSE upgrade for streams), the primary networked transport. - stdio, the local one-liner for Claude Code and Desktop.
- Unix socket (
/run/ados/mcp.sock), on-box in agent-mode, where local presence is the credential.
Safety model
- A token carries scopes (
read,safe_write,admin,flight,destructive,secret_read). Flight and destructive are off by default. - Every write is checked against the token scope, a per-tool safety class, and, for flight or destructive actions, a typed confirmation and an operator-present signal.
- Annotations (
readOnlyHint,destructiveHint) are advertised honestly but are hints, never the enforcement point. The server enforces. - Every call produces one redacted audit event.
Develop
pnpm install
pnpm typecheck
pnpm lint
pnpm test
pnpm build
The connector is a thin shim, so it holds no telemetry buffers and runs no inference. It calls the platform's existing interfaces and presents them through MCP.
Contributing
Issues and pull requests are welcome. The repository is developed in the open under the MIT license. Please keep contributions technical and free of any private or third-party content.
License
MIT. See LICENSE. Built for civilian use.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。