px4-ulog-mcp
MCP server for inspecting PX4 ULog flight logs, enabling AI assistants to understand, debug, and reason about drone systems.
README
<p align="center"> <img src="docs/assets/banner.jpg" alt="AI Drone Toolkit by Robotto" width="100%" /> </p>
AI Drone Toolkit
MCP servers and tools for intelligent drone development (PX4 and beyond). The toolkit gives AI assistants (Cursor, Claude, etc.) structured tools to understand, debug, and reason about drone systems — starting with PX4 flight logs and simulation-only PX4 SITL command.
This is a uv workspace monorepo: shared logic lives in a core package, and each tool is a thin, independently usable layer on top of it.
Repository layout
| Path | Description |
|---|---|
packages/robotto-drone-core |
Shared, MCP-agnostic parsers, safety checks, and coordinate-frame utilities. |
tools/px4-ulog-mcp |
MCP server for inspecting PX4 ULog (.ulg) flight logs. |
tools/px4-sitl-mcp |
Simulation-only MCP server for commanding PX4 SITL with safety checks. |
docs/ |
Project-wide documentation. |
examples/ |
Runnable examples and demos. |
Getting started
Requires Python 3.12+ and uv.
git clone https://github.com/robotto-xyz/ai-drone-toolkit.git
cd ai-drone-toolkit
uv sync --all-packages
This sets up a single virtual environment for the whole workspace. From there, work on any member, e.g. run the PX4 ULog MCP server or the PX4 SITL command server:
uv run px4-ulog-mcp
uv run px4-sitl-mcp
Examples
Runnable scripts live in examples/ and drive the shared core
package or tool layers directly:
# Summarize a flight log (defaults to the PX4 sample fixture):
uv run python examples/analyze_ulog.py /absolute/path/to/flight.ulg
# Flag warnings/errors and the flight-mode timeline:
uv run python examples/flight_health_check.py /absolute/path/to/flight.ulg
# Requires a running PX4 SITL instance:
uv run python examples/sitl_connect_state.py
uv run python examples/sitl_fly_survey.py
See examples/README.md for details.
Documentation
docs/— architecture and project-wide guides.packages/robotto-drone-core/README.md— shared parsing, safety, and frame-conversion APIs.tools/px4-ulog-mcp/README.md— install, wire into your editor, and analyze flight logs.tools/px4-sitl-mcp/README.md— connect to PX4 SITL, inspect state, and issue simulation-only commands.
License
MIT.
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。