Request for NIM API Rate Limit Increase (40 → 200 RPM) forParallel Multi-Agent Workflow

AI & Data Science NVIDIA NIM Access/Accounts
1

I’m an individual developer building a personal, non-commercial multi-agent coding assistant — an orchestrator that delegates to specialized sub-agents (planning, research, coding, and parallel web-search ‘scouts’). It runs on an open-source agent harness with NIM as the model backend, which is genuinely the best free option I’ve found for this.

The architecture is inherently bursty: one user request can spawn several agents that call NIM concurrently, producing 10–30 requests within a single minute. At the 40 RPM free-tier limit this regularly returns HTTP 429, which interrupts long-horizon runs and makes the system unreliable to test.

I’d like to request an increase to 200 RPM (or the next available individual-developer tier) so I can keep iterating.

Account details:

Thank you for making this catalog available to individual developers — it’s been invaluable for learning.

@TomNVIDIA @sophwats — tagging for visibility.

2

Okay i’m sorry to tell you this but there’s tons of Local models you can run for this… you’re little Open source Agentic Stuff has Traumatized Kimi K2.6… now it’s not even working right anymore.

3

Yeah… so many traumatic moments exploring AI. I tried running local models and that traumatized my poor laptop. Then I subscribed to Kimi’s Allegretto plan, which ran out in just 10 days — that traumatized me. Now I need to find non-traumatic ways to explore… 😉

4

you have Nemotron or Llama 32B on here.. and there’s also Nemoclaw on the build page..}
Also API Rate Limit Increase is NOT granted by requesting it here - #3