Hi NVIDIA NIM Support Team,
I am a personal developer building multi-agent workflows using the Hermes Agent framework, with NVIDIA NIM as the primary inference backend.
A single complex agentic task (research, tool-calling, code execution) often requires 20-50 sequential API calls. With the current 40 RPM limit, I hit 429 errors within minutes, completely blocking development.
Account Email: mc0121@126.com
Organization: Personal
Current Limit: 40 RPM
Requested Limit: 200 RPM
This request is for personal, non-commercial R&D only. I already implement exponential backoff and responsible usage.
An increase to 200 RPM would allow me to conduct meaningful, uninterrupted multi-agent tests and fully evaluate NIM for autonomous agent applications.
Thank you for considering this request.