Product
One API. Two products.
Urdu speech-to-text, and English, Arabic & Urdu text-to-speech — built for Pakistan and the Middle East.
Text-to-Speech
Turn text into natural speech in English, Arabic and Urdu. Real-time, streaming, or batch.
- English
- Arabic
- Urdu
- Real-time API
- Streaming
- Batch synthesis
Speech-to-Text
Transcribe Urdu audio into clean, structured text. Built for real calls, not studio recordings.
- Urdu transcription
- Batch transcription
- Streaming transcription (chunked)
Core Features
Everything the API gives you.
Audio Formats
Output in whatever your stack expects.
Model Types
Pick the model that fits the job.
Realtime TTS
Optimised for live conversations — lowest time to first audio.
High-Quality TTS
Higher-fidelity synthesis for pre-recorded content and batch jobs.
Urdu STT
Transcription tuned for real Pakistani speech, not studio recordings.
Custom Enterprise
Fine-tuned models and custom voices for dedicated deployments.
Text & Speech Processing
Handles the messy parts of real language.
TTS
- Urdu numbers, dates & currency
- Phone number normalisation
- Arabic diacritics
- English–Urdu code-switching
- English–Arabic code-switching
STT
- Streaming transcription via chunking
- Word-level timestamps (Business & Enterprise)
API Performance
<150ms
First-audio latency (TTS)
<200ms
Transcription latency (STT)
0.3×
Real-time factor
1000+
Concurrency
STT + TTS
Streaming
Built For
Where teams put this to work.
Customer Support
Answer every customer in their own language, day or night.
AI Receptionist
Book appointments and route calls with no human on the line.
IVR
Replace robotic menus with a voice people actually understand.
Call Transcription & Analytics
Turn Urdu calls into searchable, structured text.
Announcements
Clear, natural announcements in the language your audience speaks.
Storytelling
Give every story a natural voice in Urdu, Arabic or English.