CallRolin V2 · نموذج habibi-v1 · صوت عربي بالذكاء الاصطناعي

النص العربي
يتحوّل إلى صوتٍ حقيقي

تحويلٌ احترافي للنصوص العربية إلى كلامٍ طبيعي. اكتب أي نص واسمعه بصوتٍ واضحٍ وسلس.

🌏 الخادم في آسيا · Server in Asia زمن الشبكة · Network … يُقاس زمن الوصول مباشرةً من موقعك. الطلبات من قارّةٍ أخرى تضيف عادةً ‎150–300ms.

النص العربي 0 / 2000

Pronunciation variety 0.120

Speaking speed 1.0×

Slow 0.5× → Normal 1.0× → Fast 2.0×

Rhythm variation 0.500

—

First byte · TTFB

—

Server latency

—

Round trip

—

Characters

—

Audio length

Latest output First byte — ms · total — ms · — chars

0:00 / 0:00 ⬇

Generation history

لا يوجد صوت حتى الآن

⚡

Concurrent Load Test

Test multiple simultaneous requests for latency & throughput

Concurrent requests:

Running…

—

Avg latency

—

Min latency

—

Max latency

—

Success rate

Overview

CallRolin V2 Arabic TTS (Beta) provides neural Arabic text-to-speech synthesis. Send Arabic text and receive high-quality WAV audio.

Base URL

All API requests use JSON request bodies and return audio/wav responses.

Latency & Regions

The inference server is hosted in Asia. Round-trip network time therefore depends heavily on where your request originates.

Time to first byte (TTFB) is the interval between sending the request and receiving the first byte of audio. The Playground measures it live by streaming the response — it is the most honest signal of how responsive synthesis feels, because it excludes full-audio download time.

Client region	Typical network RTT	Feels like
Same region (Asia)	~10–60 ms	Instant
Middle East / Gulf	~90–160 ms	Fast
Europe	~150–260 ms	Noticeable
Americas	~230–350 ms	Cross-continent lag

These are network round-trips only, added on top of server inference time (X-Latency-Ms). For latency-sensitive deployments in another continent, an on-premise or in-region node removes this gap entirely.

Authentication

Include your API key in the X-API-Key header with every request.

Your API Key

Enter your API key above — all code examples update automatically.

Header	Value	Required
`X-API-Key`	Your secret API key (prefix: hbbi-)	Required
`Content-Type`	`application/json`	Required

POST /v1/tts/synthesize

Convert Arabic text to speech. Returns WAV audio.

POST

Request Body

Parameter	Type	Required	Default	Description
`model`	string	Required	—	Must be `habibi-v1`
`text`	string	Required	—	Arabic text. Max 2000 chars.
`noise_scale`	float	Optional	0.12	Pronunciation variety. Range 0–1.
`length_scale`	float	Optional	1.0	Duration multiplier. 0.5=2× faster, 2.0=2× slower.
`noise_scale_w`	float	Optional	0.5	Rhythm variation. Range 0–1.
`denoise`	bool	Optional	true	Apply denoising to output audio.

Response Headers

Header	Description
`X-Request-Id`	Unique request ID
`X-Model`	`habibi-v1`
`X-Latency-Ms`	Server inference time in ms (excludes network)
`X-Text-Chars`	Input character count

Code Examples

Models

Model ID	Language	Sample Rate	Region	Description
`habibi-v1`	Arabic (ar)	22,050 Hz	Asia	Neural Arabic TTS with sentence-level chunking and denoising. Gulf dialect optimized.

Error Codes

HTTP Status	Error Code	Cause
401	missing_api_key	No API key in request headers
401	invalid_api_key	Key is incorrect or deactivated
400	invalid_model	Model ID not recognized
429	rate_limit_exceeded	Max 3 concurrent API requests
500	inference_error	Internal model error

النص العربييتحوّل إلى صوتٍ حقيقي