Declined
0
n8n MCP server tool descriptions cause excessive token usage in Claude, hitting conversation limits quickly
Source: n8n-io/n8n#17144 · opened by @mreaves15
Bug Description
The n8n MCP server (@n8n/n8n-mcp) sends extremely verbose tool descriptions in every exchange with Claude Desktop, causing conversations to hit token limits after just 2-3 messages. This makes the integration nearly unusable.
To Reproduce
1.Install n8n MCP server via Claude Desktop config:
json
"n8n-mcp": {
"command": "npx",
"args": ["-y", "@n8n/n8n-mcp"]
}
2. Start a conversation with Claude about n8n
3. Observe that conversation limits are reached within 10-20 messages or less
Expected behavior
Tool descriptions should be concise (under 200 characters)
Detailed documentation should be available via a separate endpoint or command
Users should be able to have extended conversations without hitting token limits
Operating System
windows
n8n Version
1.100.1
Node.js Version
v22.16.0
Database
SQLite (default)
Execution mode
main (default)
The n8n MCP server (@n8n/n8n-mcp) sends extremely verbose tool descriptions in every exchange with Claude Desktop, causing conversations to hit token limits after just 2-3 messages. This makes the integration nearly unusable.
To Reproduce
1.Install n8n MCP server via Claude Desktop config:
json
"n8n-mcp": {
"command": "npx",
"args": ["-y", "@n8n/n8n-mcp"]
}
2. Start a conversation with Claude about n8n
3. Observe that conversation limits are reached within 10-20 messages or less
Expected behavior
Tool descriptions should be concise (under 200 characters)
Detailed documentation should be available via a separate endpoint or command
Users should be able to have extended conversations without hitting token limits
Operating System
windows
n8n Version
1.100.1
Node.js Version
v22.16.0
Database
SQLite (default)
Execution mode
main (default)
No pledges yet. Be the first to back this.
Comments
Similar requests
MCP tool calls fail in queue mode: `Tool node "<name>" does not have supplyData method` + `Worker tool execution timeout`
0 votes · 0 comments
AI Agent: return tool-schema validation failures to the model as recoverable errors instead of terminating the execution
0 votes · 0 comments
Anthropic Chat Model: expose "strict": true for client tool definitions
0 votes · 0 comments
Feature Suggestion: Optional token metering & paid API key support via `neuforge-pay`
0 votes · 0 comments
Feature Suggestion: Optional token metering & paid API key support via `neuforge-pay`
0 votes · 0 comments
No comments yet.