Hello NVIDIA Support Team,
I’m reaching out to request a bump in my NIM API rate limit. I’m currently stuck at the default 40 RPM, and I’m hoping to get upgraded to 200 RPM (or whatever the next standard tier is for individual developers).
Account Email: abdelahadmason@gmail.com
API Key ID (Last 4 chars): IZdf
To give you some context on what I’m building, I’m prototyping a completely free, DIY terminal user interface (TUI). It basically functions as a super-powered CLI coding assistant driven by an autonomous agent swarm.
The issue I’m running into is how these agent swarms actually operate under the hood. When I type a command into my terminal, my main orchestrator model has to digest a massive context window—I’m currently pushing the 256k token limit to ingest entire codebases. Once it understands the prompt, it instantly fires off simultaneous requests to a bunch of specialized sub-agents, like file readers and syntax checkers.
Because the orchestrator and all the sub-agents are hitting the API at the exact same time, I blow past that 40 RPM limit in a matter of seconds.
Right now, I’m constantly getting bombarded with HTTP 429 (Too Many Requests) errors. It completely breaks the TUI experience. The only way I can get it to run is by heavily throttling the concurrent workers, which ruins the speed and whole point of having a terminal assistant in the first place.
If I could get that 200 RPM bump, it would let me run real-world testing without my script crashing every time I launch a complex multi-agent task. Just to be clear, this is purely a personal development project, not a massive commercial application driving production traffic.
Thanks for looking into this!
Best, *
Mason*