How do I use Nova-3 from inside Iran?

Create a Hormouz account and generate an API key. Hormouz gives you a stable route to the model from inside Iran, with no foreign credit card or circumvention tooling needed.

How do I call Nova-3?

Hormouz serves Nova-3 on two endpoints. The OpenAI-compatible POST /v1/audio/transcriptions takes the file as the multipart field "file", or as JSON input_audio with base64 data and format, and returns {text, usage} in the same response. POST /v1/predictions handles long recordings in the background: input.audio is an https URL or a data: URI, input.language is optional. For Persian, set the language to "fa". Code written for OpenAI's SDK works by changing only the base URL.

How do I get Persian transcripts from Nova-3?

Set the language to "fa": the language field on /v1/audio/transcriptions, or input.language on /v1/predictions. Without it, Persian audio can come back as an empty transcript.

Deepgram: Nova-3: audio model

Nova-3 is Deepgram's general-purpose speech-to-text model, with monolingual and multilingual transcription. Reachable from inside Iran through Hormouz.

Pricing:: $0.0052/minute of audio
Type:: audio

Audio

Nova-3 is a general-purpose speech-to-text model from Deepgram, with support for both monolingual and multilingual transcription.

Hormouz serves Nova-3 on two endpoints: the OpenAI-compatible /v1/audio/transcriptions, which returns the text in the same response, and /v1/predictions for long recordings processed in the background. For Persian, set the language to "fa": in our tests Persian worked only with that setting, and without it the transcript could come back empty.

Billing is per minute of audio; the current rate is in the price table on this page, and you pay only for what you use. Hormouz provides stable access to it from inside Iran, with no foreign credit card or circumvention tooling required.

Deepgram Nova-3 general-purpose speech-to-text model with monolingual and multilingual transcription support.

Quickstart

curl https://api.hormouz.net/v1/audio/transcriptions \
  -H "Authorization: Bearer ng-YOUR_API_KEY" \
  -F model=deepgram/nova-3 \
  -F file=@audio.mp3

Specs

Vendor	Deepgram
Type	Audio
Context window	—
Pricing	$0.0052/minute of audio
API	transcriptions

Code example

cURL

curl https://api.hormouz.net/v1/audio/transcriptions \
  -H "Authorization: Bearer ng-YOUR_API_KEY" \
  -F model=deepgram/nova-3 \
  -F file=@audio.mp3

Python

from openai import OpenAI
client = OpenAI(api_key="ng-YOUR_API_KEY", base_url="https://api.hormouz.net/v1")
with open("audio.mp3", "rb") as f:
    tr = client.audio.transcriptions.create(model="deepgram/nova-3", file=f)
print(tr.text)

JavaScript

import OpenAI from "openai";
import fs from "node:fs";
const client = new OpenAI({ apiKey: "ng-YOUR_API_KEY", baseURL: "https://api.hormouz.net/v1" });
const tr = await client.audio.transcriptions.create({
  model: "deepgram/nova-3",
  file: fs.createReadStream("audio.mp3"),
});
console.log(tr.text);

Buy the Deepgram: Nova-3 API, paid in rial

Buy credit with an Iranian bank card; no foreign card or account needed.
One Hormouz key reaches Deepgram: Nova-3 and 600+ other models, and you pay only for what you use.
No VPN and no risk of a banned account; just change the base URL to api.hormouz.net/v1.

Create an account

Use via Hormouz

Connect to this model and 600+ others with one API key, then see usage down to each request.

Request company access

FAQ

How much does Nova-3 cost?: Billing is per minute of audio; the current rate is in the price table on this page, and you pay only for what you use.
How do I use Nova-3 from inside Iran?: Create a Hormouz account and generate an API key. Hormouz gives you a stable route to the model from inside Iran, with no foreign credit card or circumvention tooling needed.
How do I call Nova-3?: Hormouz serves Nova-3 on two endpoints. The OpenAI-compatible POST /v1/audio/transcriptions takes the file as the multipart field "file", or as JSON input_audio with base64 data and format, and returns {text, usage} in the same response. POST /v1/predictions handles long recordings in the background: input.audio is an https URL or a data: URI, input.language is optional. For Persian, set the language to "fa". Code written for OpenAI's SDK works by changing only the base URL.
How do I get Persian transcripts from Nova-3?: Set the language to "fa": the language field on /v1/audio/transcriptions, or input.language on /v1/predictions. Without it, Persian audio can come back as an empty transcript.

Updated 2026-10-13