enes (ekkovoice)
04/10/2026, 11:05 AMAI Evy
04/10/2026, 11:05 AMAI Evy
04/10/2026, 11:06 AMnl-NL (Dutch - Netherlands) or nl-BE (Dutch - Belgium) for the language setting [(1) ]() [(2) ]() . This optimizes speech recognition specifically for Dutch.
2. Use accurate STT mode
Set stt_mode to "accurate" rather than "fast" to prioritize transcription accuracy over speed . This is particularly important for challenging audio conditions like hands-free car calls.
3. Add boosted keywords
Use the boosted_keywords parameter to bias the transcriber model toward specific Dutch words, names, brands, or phrases that are commonly used in your calls [(1) ]() [(2) ]() [(3) ]() . For example:
json
["retell", "kroger"]AI Evy
04/10/2026, 11:06 AMdenoising_mode setting to "noise-cancellation" to filter out ambient sounds .
These configurations should significantly improve Dutch speech recognition consistency, even in challenging hands-free car scenarios where audio quality may be compromised by background noise and distance from the microphone [(4) ]() [(5) ]() .AI Evy
04/10/2026, 11:06 AMAI Evy
04/10/2026, 11:06 AM