How to fix MCP tool overload

In partnership with

The Simplest Way to Create and Launch AI Agents and Apps

You know that AI can help you automate your work, but you just don't know how to get started.

With Lindy, you can build AI agents and apps in minutes simply by describing what you want in plain English.

→ "Create a booking platform for my business."
→ "Automate my sales outreach."
→ "Create a weekly summary about each employee's performance and send it as an email."

From inbound lead qualification to AI-powered customer support and full-blown apps, Lindy has hundreds of agents that are ready to work for you 24/7/365.

Stop doing repetitive tasks manually. Let Lindy automate workflows, save time, and grow your business

Get $20 Worth of Free Credits

Over the past few months, developers experimenting with MCP servers have encountered the same challenges. Agents start strong but soon start drifting. It doesn’t make the expected tool calls. The root cause here, though, is loading too many tools at once.

Once an MCP server carries 30, 40, or 50 exposed tools, we soon see the performance tank. Cursor caps MCP tools at 40 tools for stability and Copilot limits to 128. Claude Code also performs worse 15–20 active tools

When the model has to parse dozens of similar tool names and parameters, it loses focus. It’s literally trying to find the needle in a haystack.

So how to fix it?

1. MCP with Multiple Configurations

You can add and manage MCP servers separately by scope local, project shared, or user wide and load multiple configuration files as needed.

This keeps your sessions lightweight, organized and focused by avoiding loading unnecessary MCP tools. Perfect for managing different environments or permissions. Instead of one massive JSON, you can load only the configs you need.

// github_read.json
{
  "mcpServers": {
    "github-read": {
      "url": "https://api.githubcopilot.com/mcp/",
      "authorization_token": "GITHUB_READ_ONLY_TOKEN"
    }
  }
}

// github_write.json
{
  "mcpServers": {
    "github-write": {
      "url": "https://api.githubcopilot.com/mcp/",
      "authorization_token": "GITHUB_WRITE_TOKEN"
    }
  }
}

You can load both configurations simultaneously in Claude Code, or load only the configurations necessary for the session.

claude --mcp-config github_read.json github_write.json 

2. Workflow profiles

Keep tool sets separate per workflow by defining distinct MCP profiles per role or task agent, then load only the profile needed for the current session to avoid dragging unrelated servers into context.

Scopes tools by agents (e.g., coding, research, ops) so each session loads a minimal, relevant set rather than a global superset, reducing manifest size and selection noise.

3. Avoid one-to-one API to MCP conversion

If you have your own MCP servers, try combining multiple APIs that achieve a vertical task instead of just wrapping individual API calls. You can see this on how to do this for more details.

Takeaway

The lesson is simple. Optimize for context, not quantity.

Can you take a moment to rate this edition? It helps me cater the right content to you.

📱 Stay connected: Follow me on LinkedIn and Twitter/X for daily AI tool breakdowns and quick wins