Arabic AI Voice

Noura — free Kuwaiti Arabic text to speech

Noura is a free female Kuwaiti Arabic AI voice — a Microsoft neural voice. Type your text below to hear it read aloud, then download the audio. No account needed.

Female Kuwaiti accent Global voice Free

Make Noura say your words

Voice details

Language
Arabic
Accent
Kuwaiti
Gender
Female
Type
Global · Microsoft neural
Download
MP3
Speed
Natural pace
Voice ID
ar-KW-NouraNeural
Price
Free

How Noura sounds — measured

Median pitch

214 Hz

In the middle of 16 female Arabic voices

Intonation range

10.5 semitones

Close to typical for its group

Tone (spectral centroid)

1,657 Hz

Balanced tone for its group

Speaking rate

9.5 characters/s

In the middle of 32 Arabic voices on the same passage

Pitch across the sample passage dashed line = Noura’s median

How long your audio will be at 9.5 characters/s

1,000 characters
≈ 1 min 45 s
5,000 characters (free limit per render)
≈ 9 min
  • ↔Compared with Fatima, Noura has a similar pitch (214 vs 214 Hz) and faster delivery (9.5 vs 8.6 characters/s). Listen to Fatima →
  • ↔Compared with Aysha, Noura has a similar pitch (214 vs 214 Hz), livelier intonation (10.5 vs 2.7 semitones) and slower delivery (9.5 vs 11.2 characters/s). Listen to Aysha →

Measured by SpeakScribe, October 2026 · How we measure · Every Arabic voice side by side

Use Noura in a conversation

With Multi-voice, Noura can read one character while Shakir (a male Arabic voice) reads another — the whole dialogue comes back as one MP3 with even volume between speakers. A free account can make 3 two-speaker dialogues a month; paid plans allow up to 4 speakers on Starter.

Try a dialogue with Noura →

Formats, limits and help

Noura: quick answers

Measured from its audio, Noura has a median pitch 214 Hz (in the middle of 16 female Arabic voices); intonation range 10.5 semitones (close to typical for its group); tone (spectral centroid) 1,657 Hz (balanced tone for its group).
About 1 min 45 s. Noura reads about 9.5 characters a second, spaces included (measured on its sample passage), so the free limit of 5,000 characters per render is about 9 minutes of audio. Numbers, names and long pauses make it longer.