File size: 1,099 Bytes
28f7e95
a66d269
 
 
 
28f7e95
 
a66d269
28f7e95
 
a66d269
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
---
title: MONAI MCP Service
emoji: 🔧
colorFrom: blue
colorTo: indigo
sdk: docker
pinned: false
license: mit
---

# MONAI MCP Service

This deployment package wraps MONAI capabilities as MCP tools using FastMCP.
It supports local `stdio` usage and HTTP MCP endpoint deployment for Docker/HuggingFace Spaces.

## Available Tools

- `health_check`
- `list_modules`
- `list_symbols`
- `list_network_architectures`
- `create_network`
- `build_transform_pipeline`
- `call_function`
- `create_instance`

Detailed tool docs are in `MONAI/mcp_output/README_MCP.md`.

## Local stdio (Claude Desktop / CLI)

```bash
cd MONAI/mcp_output
python mcp_plugin/main.py
```

or:

```bash
MCP_TRANSPORT=stdio python start_mcp.py
```

## HTTP MCP endpoint (Docker/HF Spaces)

The Docker entrypoint launches FastMCP directly from:

- `python MONAI/mcp_output/start_mcp.py`

Runtime env in container:

- `MCP_TRANSPORT=http`
- `MCP_PORT=7860`

MCP clients should connect to:

- `https://<your-host>/mcp`

## Build and Run Docker

```bash
./run_docker.sh
```

On Windows PowerShell:

```powershell
./run_docker.ps1
```