Ask HN: How do you manage multiple MCP servers in Claude Code?

2 d_sc 3 9/3/2025, 10:36:42 AM
I have a handful of MCP servers that I really like to use: Sentry, Shortcut, Playwright. There’s a few others that I’d like to use.

My problem, even with a couple of MCP servers installed the number of tools loaded easily surpasses the 40 tool threshold and I see warnings in Cursor and Claude Code.

My current solution is to manually turn on tools within each MCP server. The challenge with this is that it 1) takes a decent amount of time to enable/disable individual tools 2) you need to understand what each tool does and whether disabling one will impact the functionality of another tool and 3) different types of tasks require different sets of tools enabled (ie researching vs. speccing vs. testing).

How are others managing what tools are enabled & disabled?

Comments (3)

pmig · 13h ago
I am also running into similar problems, constantly seeing Claude Code warnings that my tools take up a lot of the context window. There are some projects out there that try help you build a unified MCP server (aci.dev, ibm-contextforge, hyprmcp.com[1]). Also the MCP protocol supports a notifications[2] for "dynamically adding tools", but most clients don't support them. A "smart" middle layer still needs to be created.

[1]: Disclosure I am working at https://hyprmcp. Our gateway is fully open source: https://github.com/hyprmcp/mcp-gateway

[2]: https://modelcontextprotocol.io/docs/learn/architecture#noti...

thsvrrck · 12h ago
I just use markdown to keep track of all my MCP connections. Makes it easy to see which tools are enabled for each workflow, and switching setups is quick. Anyone else doing something similar?
__warlord__ · 12h ago
* Use subagents with access to dedicated tools. * Build slash commands with high level directives to use those sub agents in a structured way