Skip to main content
Glama

embodied-nav-mcp

一个语言模型,可以环顾房间,走向它命名的物体,数出所见之物,并在完成时停下。

模型决定做什么。传感器决定物体在哪里

工具

工具

功能

get_view

来自360°相机的四个透视面,加上位姿

ground

指代表达式 → 像素框 → 激光雷达米数 → 转换器吸附的路点

drive_to

发布一个地图坐标系下的 (x, y) 并等待到达或稳定

stop

停车,使规划器不再追逐最后一个路点

parse_numerical

将计数问题拆分为目标/锚点

count_in_view

仅统计此视角;每个实例都通过激光雷达提升

commit_count

按位置聚类观察;发布整数

publish_object

最后一个 ground 的激光雷达 AABB;发布3D标记

commit_count 不从模型获取数字。四个枕头的两个视角变成4,而不是8。

几何部分来自 Team Xiao Hei 的 feat/xiao-hei-instruction-merge 的 Type 3 栈:尺寸门、盲锥处理、转换器稳定位姿、退出时停车。

安装

uv sync --extra dev
export ANTHROPIC_API_KEY=...
uv run pytest

Python 3.11+。真实机器人/仿真需要 ROS Jazzy 和自主栈。测试使用假机器人,两者都不需要。

尝试

# Dry run, no ROS
uv run embodied-mcp-agent --fake "How many blue chairs are between the table and the wall?"
uv run embodied-mcp-agent --fake "Find the potted plant on the kitchen island closest to the fridge."
uv run embodied-mcp-agent --fake "Take the path near the window to the fridge."

在机器人(或挑战容器)上,栈已启动时:

uv run embodied-mcp-challenge
# evaluators still run: ros2 launch dummy_vlm dummy_vlm.launch

embodied-mcp --attach-ros 是通过 stdio MCP 的相同工具。挑战节点在进程内调用它们:机器人容器没有用于远程 MCP 连接器的公共 URL。

CMU VLN 挑战

即插即用适配器:examples/cmu_vln/。包名 dummy_vlm 保持不变。

类型

工具

ROS 输出

数值

parse_numerical, count_in_view, commit_count

/numerical_response

物体指代

ground, publish_object

/selected_object_marker

指令跟随

get_view, ground, drive_to, stop

/way_point_with_heading

允许的传感器:/camera/image/registered_scan/terrain_map/terrain_map_ext/state_estimation。到达状态是 /state_estimation,而不是 /way_point_reached。每个进程一个问题。

配置

变量

默认值

ANTHROPIC_API_KEY

必需

EMBODIED_MCP_MODEL

claude-opus-5

主机和像素接地/计数模型

EMBODIED_MCP_BUDGET_S

540

从进程启动起的秒数

XIAO_HEI_IMAGE_TOPIC

/camera/image

相机话题

许可证

MIT。内置的几何和 ROS I/O 来自 Team Xiao Hei;参见 NOTICE

-
license - not tested
Not graded
quality - not tested
C
maintenance

Maintenance

Maintainers
Response time
Release cycle
Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/aryanmangal769/embodied-nav-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server