Spontaneous Conversational Sundanese Speech Dataset — Consented, Transcribed (West Java, Indonesia)

Real Sundanese, spoken the way it is actually lived. This dataset contains spontaneous, unscripted conversations in Sundanese (Basa Sunda) — the language of ~42 million people in West Java, Indonesia, and one of the most under-represented major languages in commercial speech AI. WHAT'S INSIDE • Recordings 48 kHz / 16-bit WAV, quiet indoor environments • Verbatim transcripts with timestamps for every segment • Speaker metadata: age band, gender, domicile, dialectal area • Per-recording consent forms and provenance documentation — every speaker consented in writing, every hour traceable WHY IT'S RARE There is effectively no commercial supply of consented, transcribed spontaneous Sundanese speech. Our corpus work in this language has been supported and validated through Mozilla's community funding program — this listing is its premium, production-grade continuation. QUALITY CONTROLS • SNR-checked recordings (target ≥ 25 dB), no clipping • Human QC pass on audio + transcript alignment • Documented collection SOP; consistent device & room standards SUITABLE FOR • ASR / speech recognition for low-resource languages • Conversational voice AI and call-center speech for West Java • Speaker diarization, language ID, sociolinguistic research • Foundation-model multilingual speech pretraining LICENSE & PRICING (USD) • Hourly license, non-exclusive, commercial use: 1–5 hours: $90/hr · 5–15 hours: $80/hr · 15+ hours: custom • Minimum order: $250 • Exclusive full-rights buyout available on