NewEmirati & Saudi voices are live.Try them →
Voice Stack by Xpin AI

Don’t type it,
say it.

Voice Stack turns plain text into natural, polished speech — studio-grade voices in English and Arabic — Emirati, Saudi and Standard — rendered 40× faster than realtime, from a web studio or one API call.

Web studio · HTTP API · WebSocket

creators, devs & support teams
Text to speech

“Take a slow breath in… and let it go. You are exactly where you need to be.”

Bilal · warmEnglish1.0× speed
0:09
Rendered in 0.21sDownload .wav ↓
English•Arabic•Emirati•Saudi•الإنجليزية•العربية•الإماراتية•السعودية•English•Arabic•Emirati•Saudi•الإنجليزية•العربية•الإماراتية•السعودية•
The platform

Every dialect, any voice.

English and Arabic from one model. Switch between Emirati, Saudi and Standard Arabic per request, pick a voice from the library — or clone your own from a short clip and use it everywhere.

BEnglish

Bilal

Male · Warm · EN

NEmirati

Noor

Female · Calm · AR

FSaudi

Fahad

Male · Deep · AR

MArabic

Maryam

Female · Crisp · AR

NarratorWarm hostNewsreaderMeditationYour voice +
Browse the voice library
Giving a voice to
podcastsagentsaudiobooksgameslearningproductssupportadsappspodcastsagentsaudiobooksgameslearningproductssupportadsapps

40× faster than realtime

A voice that finally keeps up. Render a full chapter while a studio session is still setting up the microphone — then regenerate a line as fast as you can rephrase it.

40×
faster than realtime
RTF ≈ 0.025 on a single GPU.
0.4s
to first audio
Streaming starts before you finish typing.
EN · AR
English & Arabic
Emirati, Saudi and Standard Arabic dialects.
24 kHz
studio audio
Clean WAV or raw PCM frames.
Studio session
hours

per finished chapter — booking, takes, retakes, editing.

Voice Stack
seconds (RTF ≈ 0.025)

GPU-rendered narration, ready to stream before you finish your coffee. First audio in about 0.4 seconds.

Everything a voice needs

Cloning, streaming, languages and controls — one product, one key, the same voices everywhere.

Voice cloning

One clip. Your voice, everywhere.

A short reference recording becomes a reusable voice — no retraining, instantly available by voice ID in the studio and the API.

NarratorBilal · ENWarm hostNoor · ARYour voice +
Realtime streaming

Playback starts before rendering ends.

Chunked WAV for a plain <audio> tag, raw PCM frames over WebSocket for agents and live products.

live
English & Arabic

Emirati, Saudi & Standard

Native-sounding English and Arabic — switch dialect per request.

Web studio

Write, listen, export

A browser workspace for scripts, voices and history — no setup, nothing to install.

SpeedLanguageHistory
Fine control

Speed, tone, style

Describe the delivery — gender, age, accent, pitch — and set the pace per request.

Speed
Warmth
Pitch
Developers

Voice Stack, wherever you ship

Use the web studio, or wire it straight into your product with an API key. Same voices, same speed, everywhere.

  1. 1
    Get a key
    Sign in and create an API key from your account.
  2. 2
    POST some text
    One JSON request returns a finished 24 kHz WAV.
  3. 3
    Stream it
    Switch to /stream or the WebSocket for first audio in ~0.4s.
POST /v1/tts
curl -X POST https://voice-api.xpin.ae/v1/tts \
  -H "Authorization: Bearer $KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "مرحباً، أهلاً بك في فويس ستاك",
    "language": "afb",
    "voice": "noor",
    "speed": 1.0
  }' \
  -o hello.wav
→ 200 OK · audio/wav · X-Duration-Seconds: 3.8 · rendered in 0.11s
POST /v1/tts
Full WAV
POST /v1/tts/stream
Chunked WAV
WS /v1/tts/ws
PCM frames

From text to talking

↗

One request turns a paragraph into finished narration.

~0.4s to first audio

↗

Streaming starts almost before you hit enter.

English & Arabic, one API

↗

Emirati, Saudi and Standard dialects — no extra contracts.

24 kHz studio audio

↗

Clean WAV or raw PCM, ready for production.

Start speaking.

Effortless voice for everything you write — in the studio today, in your product tomorrow.