Rural Women ASR V2: 10K-100K Hindi and Bhojpuri Speech Utterances

Rural Women ASR V2 provides between 10,000 and 100,000 audio utterances of Hindi and Bhojpuri speech recorded from rural women in India. Developed by ai4bharat for the Recognizing Every Voice initiative, the collection was last updated in November 2025. It captures speech from diverse age groups, regions, and socio-economic backgrounds to improve inclusive speech technology.

Use Cases

• Fine-tuning ASR models for the Bhojpuri language using audio-text pairs
• Benchmarking speech-to-text accuracy specifically for rural women speakers
• Linguistic analysis of Hindi dialectal variations across different rural Indian regions

Strengths

• 10,000 to 100,000 record count
• Includes underrepresented Bhojpuri language data
• Targets rural women speakers to address demographic bias in ASR

Limitations

• Limited to two specific languages
• Potential for environmental background noise given the rural collection setting

Provenance

Source

ai4bharat

Collection Method

Recorded from rural women speakers across diverse regions

Freshness

Last updated November 2025.

Geography

India (Rural regions)