What You'll Do
- Read short phrases in Swahili or English.
- Record each prompt clearly, retry if needed.
- Submit only when you're satisfied with the audio.
Contribute Voice Samples for ASR Model Training
Help us build high-quality Swahili & English speech data to improve automatic speech recognition for African dialects.
Recording Tips
- Work in low background noise environment with no nearby speech.
- Place the mic ~20–30 cm from your mouth.
- Read exactly as shown; speak naturally and clearly.
Quality Review
- We automatically screen for background noise, distortion, and too loud or too soft audio.
- Prompts can only be submitted once per contributor.
- Human reviewers ensure accuracy and language consistency.
Consent & Privacy
By continuing and submitting recordings, you confirm your consent to the collection of your audio, dialect metadata, and session activity strictly for speech model development and evaluation.
You can request deletion of your contributions at any time by contacting krishna.munjam@sunking.com.