Remy Reference/Models/Sound generation/Mirelo SFX 2.0

Mirelo SFX 2.0

Mirelo SFX 2.0, a sound effect model from WaveSpeed.

soundModelOverride: { model: "mirelo-sfx-2.0" }

MMirelo·served viaWaveSpeed·Released October 2026

Pricing

List price: $0.016/second
Per generation: $0.00010

Specifications

Context window: 2,000 tokens
Released: October 2026
Model string: mirelo-ai/sfx-2.0/text-to-audio

Parameters

Model options, passed in soundModelOverride.config.

Parameter	Type	Default	Range / options
`$mode` Sound from a description, or sound effects for a video.	Select	text-to-audio	text-to-audiovideo-to-audio
`video` The video to make sound for, at least a second long.	Video URL	—	—
`preserve_speech` Keep the video's own speech under the new sound.	Boolean	true
`duration` Up to 600. A video's sound stops at the video's end.	Number	10	1–600, step 1
`negative_prompt` Sounds to avoid, such as music, voices or traffic.	Text	—	—
`output_format`	Select	mp3	mp3wavaacflac

Use this model

await mindstudio.generateSound({
  text: "...",
  soundModelOverride: {
    model: "mirelo-sfx-2.0",
    config: {
      duration: 10,
      output_format: "mp3",
    },
  },
});