curl -sS -X POST "https://api.ai-coustics.io/v2/medias" \
  -H "X-API-Key: $AICOUSTICS_API_KEY" \
  -F file=@"./noisy.wav" \
  -F enhancement='{ "enhancement_model": "FINCH", "enhancement_level": 100, "loudness_target": -14, "true_peak": -1 }'
ai-coustics offers two complementary speech enhancement model families. Use this guide to choose the right model and set it in requests.
Finch 2 is our updated voice isolation model designed to remove undesired sounds (noise, reverb) while preserving the original speaker’s identity.
  • Best for: Strong background noise, heavy reverb, distant speakers, voice isolation needs
  • Strengths: Improved de-noising/de-reverb, fewer artifacts, more robust, faster and more energy‑efficient
  • Parameter: enhancement_model: "FINCH" (maps to Finch 2)
curl -sS -X POST "https://api.ai-coustics.io/v2/medias" \
  -H "X-API-Key: $AICOUSTICS_API_KEY" \
  -F file=@"./noisy.wav" \
  -F enhancement='{ "enhancement_model": "FINCH", "enhancement_level": 100, "loudness_target": -14, "true_peak": -1 }'

​
Choosing the right model

  • Pick Finch 2 when your priority is preserving the original voice while removing noise/reverb or isolating speech from mixed content.
  • Pick Lark 2 when you need to repair and enhance degraded recordings (e.g., phone, Zoom, old archives) for a polished, full‑band sound.
Using "FINCH" as the parameter selects Finch 2 automatically. For reconstructive enhancement, use "LARK_V2".
Learn more about the models: