(no title)
thembones | 9 months ago
Each MCP server still needs to handle auth differently depending on what it's connecting to. A GitHub MCP server needs GitHub tokens, a database server needs database credentials, an email server needs SMTP auth, etc. The client application now has to manage and securely store N different credential types instead of implementing N different integrations.
So yes, the protocol complexity is reduced, but the real operational headache (managing secrets, handling token refresh, dealing with different auth flows) just gets moved around rather than solved. In some ways this might actually be worse since you now have N different processes that each need their own credential management instead of one application handling it all.
This doesn't make MCP useless, but the "M x N to M + N" framing undersells how much complexity remains in the parts that actually matter for production deployments.
nilslice|9 months ago
Setting aside expected criticism about this being some middleware layer, but we’ve launched a solution to this problem:
An MCP “SSO”, where you install and auth your MCP servers into profiles (collections of servers), which we virtualize into a single MCP server with only a single OAuth flow — simplifying the experience substantially for both the user of the MCP servers and the clients connecting to them.
https://docs.mcp.run/blog/2025/05/14/mcp-sso/
soulofmischief|9 months ago
nurettin|9 months ago
cruffle_duffle|9 months ago
I think this problem is inherent to connecting to a bunch of different providers. Unless all the providers were the same company or had to register directly with a single company and then proxy through but even then now you’ve just moved the problem.
heisenbit|9 months ago
jononor|9 months ago
nilslice|9 months ago
https://docs.mcp.run/blog/2025/05/14/mcp-sso/