Remote MCP Server Instruction #8
Conversation
Add instruction about using this server as a remote server.
|
|
||
| ## Set Up as a Remote Server | ||
| If you prefer to set up this mcp server as a remote server, you can replace [__main__.py](./src/mcp_server_datahub/__main__.py) with the following: | ||
| ``` Python |
There was a problem hiding this comment.
is there any way we can make this simpler?
e.g. make something like mcp-server-datahub serve --port <my port> just work?
There was a problem hiding this comment.
If we just copy and paste the code to this main.py file, you can just run mcp-server-datahub serve --port <my port>
But I am not clear the overall direction for deploying this as a remote server , I wrote this instruction just in case someone wants to use it for some PoC.
|
|
||
| Solution: Replace the `uvx` bit of the command with the output of `which uvx`. | ||
|
|
||
| ## Set Up as a Remote Server |
There was a problem hiding this comment.
| ## Set Up as a Remote Server | |
| ## Advanced: Set Up as a Remote Server |
|
I'm tentatively adding some args as part of #17 so you can do In general, I'm a bit apprehensive about adding docs around running a remote MCP server in this way because it encourages people to have a setup where all mcp server users share the same tokens. Especially as we move towards write-oriented tools (e.g. #16), having docs on that encourage shared credentials doesn't feel like right. Note that the DataHub Cloud Remote MCP server doesn't suffer this issue, since it takes a different approach to auth https://docs.datahub.com/docs/features/feature-guides/mcp |
2 participants
Add instruction about using this server as a remote server.