Request for Access: Riva TTS Magpie

AI & Data Science Deep Learning (Training & Inference) Riva
1

Hello NVIDIA team,

NVIDIA customer service told me to post my request here.

I recently applied for access to the Riva TTS Magpie Zeroshot model, but received a message that my application “needs more information”. The message didn’t specify which details were missing.

I’m developing a on-premise voice assistant prototype on the DGX spark for internal research and educational use (non-commercial). If it doesn’t work on dgx spark, I will probably try it on a cloud account.

Could you please let me know what additional details or documentation are required to complete the access review? I’m happy to provide any information needed.

Env:
DGX Spark

Thanks

4

Hello NVIDIA Riva Team,

I am facing the same issue. I applied for the NVIDIA Riva TTS Zero-shot Early Access program (Magpie TTS Zeroshot model) and received the following response on February 22, 2026:

“Please provide an NVIDIA contact person and avoid sending requests from a personal email account.”

I am an independent developer and student working on speech AI and conversational AI projects. I do not have a direct NVIDIA contact person. I am interested in integrating the Magpie TTS Zeroshot model into my AI development work for non-commercial, research/educational purposes.

Could the NVIDIA Riva team please advise on how an independent developer without an enterprise NVIDIA contact can proceed with this access request? Is there an alternative pathway or additional information I can provide to get access?

Thank you for your time.
Kasinath R

5

skimdt - access was granted.
karthikasi225 - please apply again for access

6

can you pls grant access also to my account ? I already requested it:

7

Hello NVIDIA Riva Team,

I applied for access to the Riva TTS Magpie Zeroshot model and received the following response:

“Please provide an NVIDIA contact person and avoid sending requests from a personal email account.”

I am an independent developer and do not have an existing NVIDIA contact.

I am building a private, locally hosted real-time conversational character platform combining persistent LLM-driven characters, live avatar rendering, visual context, and streaming speech. I would like to evaluate Magpie as a low-latency local TTS alternative for distinct character voices.

Environment:
NVIDIA RTX 3090 24GB
64GB RAM
Windows-based local development system

Could you please review my existing application or advise?

Thank you.