Alvoff Inference provides cheap and reliable API for speech-to-text, text-to-speech based out of India. We aim to be the one stop solution for voice related inference. Our infrastructure is purpose-built for audio workloads, which means low latency and lower cost compared to general-purpose cloud providers. You can signup on the platform and get $5 of inference for free.
No reviews yetBe the first to leave a review for Alvoff Inference - Fast, cheap STT · TTS
Maker
📌
Hey PH, Ron this side. I am one of the devs who brought this to life. We are aiming to become one stop solution for anything voice related. We currently 4 models we provide inference for and are looking to add in more capacity soon. If you are a startup or an enterprise that is looking to get some cheap inference for your STT and TTS workloads we can help you out in getting one of the cheapest inference in the market. Just shoot a mail to support@alvoff.ai and we will get back to you right away.
Report
how does the pricing actually compare per minute to something like deepgram once you burn through that $5 credit
Deepgram's pay-as-you-go Nova-2 is $0.0043/15s = ~$1.03/hour. Their Growth tier is $0.0036/15s (~$0.86/hour).
So your pricing is roughly 30–37x cheaper than Deepgram for speech-to-text. That ₹500 (~$5.90) free credit gets about 2,458 hours of transcription — it's an enormous amount.
Report
tested the stt endpoint on a noisy podcast clip and it came back clean way faster than i expected for the price. nice to see an api like this coming out of india.
Latency on the TTS endpoint was noticeably snappier than what I was getting elsewhere, and the $5 free credit was enough to actually test a real use case. Solid option if you need voice inference without the usual hyperscaler pricing.
Report
Maker
@duyguekinnq1t Thanks for checking out the platform! Glad you liked it.
Report
Out of India with infrastructure purpose-built for audio, that's a refreshing angle. Latency on the STT side felt snappy in my quick test, and the $5 credit is enough to properly eval it.
Report
how does the latency actually compare to something like openai's whisper or elevenlabs in real world conditions, not just on paper?
Report
Voice features always felt like something only the big players could pull off, so seeing this come within reach for smaller teams is quietly exciting to me. Nice to watch that door open a little wider.
how does the pricing actually compare per minute to something like deepgram once you burn through that $5 credit
@bnyamint9fq as per claude:
Deepgram's pay-as-you-go Nova-2 is $0.0043/15s = ~$1.03/hour. Their Growth tier is $0.0036/15s (~$0.86/hour).
So your pricing is roughly 30–37x cheaper than Deepgram for speech-to-text. That ₹500 (~$5.90) free credit gets about 2,458 hours of transcription — it's an enormous amount.
tested the stt endpoint on a noisy podcast clip and it came back clean way faster than i expected for the price. nice to see an api like this coming out of india.
@salime7l8 Thanks for checking out the platform!
Latency on the TTS endpoint was noticeably snappier than what I was getting elsewhere, and the $5 free credit was enough to actually test a real use case. Solid option if you need voice inference without the usual hyperscaler pricing.
@duyguekinnq1t Thanks for checking out the platform! Glad you liked it.
Out of India with infrastructure purpose-built for audio, that's a refreshing angle. Latency on the STT side felt snappy in my quick test, and the $5 credit is enough to properly eval it.
how does the latency actually compare to something like openai's whisper or elevenlabs in real world conditions, not just on paper?
Voice features always felt like something only the big players could pull off, so seeing this come within reach for smaller teams is quietly exciting to me. Nice to watch that door open a little wider.