MCP Token Trap: Why Your AI Agent Burns 35x More Tokens Than a CLI

2 comments

• Eldodi 30 minutes ago
  This is starting to get old. Both Claude and OpenAi solved the mcp context bloat issue with on-demand tool discovery:

  - Claude tool search: https://platform.claude.com/docs/en/agents-and-tools/tool-us...

  - GPT tool search: https://developers.openai.com/api/docs/guides/tools-tool-sea...
• sharemywin 1 hour ago
  I guess I don't understand. so how does the agent know which tools to call versus an mcp?