Mirelo

Vendor: Mirelo · Official API docs: Mirelo SFX

Mirelo SFX 2.0 generates sound effects and soundscapes. Three entry points share one engine: text-to-audio builds a sound from a prompt alone (1–600 s); video-to-audio watches your clip and returns a soundtrack synced to what happens on screen; video-to-video does the same and returns the clip with the new audio muxed in. Both video modes keep existing speech by default (--preserve-speech), and every mode can return up to 4 variations per call.

Mirelo SFX 2.0 is available through the CLI, MCP server and SDK. It is not currently selectable in the AI Playground web app.

Modes: video · audio · Models: 3

Models

id	Name	Input type
mirelo-sfx-2.0-text-to-audio	Mirelo SFX 2.0 (Text to Audio)	t2a
mirelo-sfx-2.0-video-to-audio	Mirelo SFX 2.0 (Video to Audio)	v2a
mirelo-sfx-2.0-video-to-video	Mirelo SFX 2.0 (Video to Video)	v2v

CLI

# a 10-second sound effect from text, as mp3
gen-ai generate -m mirelo-sfx-2.0-text-to-audio \
  -p "heavy rain on a tin roof, distant thunder" -d 10 --format mp3 -n 1

# a soundtrack synced to a clip, returned as audio
gen-ai generate -m mirelo-sfx-2.0-video-to-audio --video ./clip.mp4 \
  -p "footsteps on gravel, wind" -n 1

# the same clip with the new audio muxed back in
gen-ai generate -m mirelo-sfx-2.0-video-to-video --video ./clip.mp4 -n 1

MCP

{ "name": "picsart_generate",
  "arguments": {
    "model": "mirelo-sfx-2.0-text-to-audio",
    "prompt": "heavy rain on a tin roof, distant thunder",
    "duration": 10,
    "count": 1
  } }

{ "name": "picsart_generate",
  "arguments": {
    "model": "mirelo-sfx-2.0-video-to-video",
    "videoUrl": "https://example.com/clip.mp4",
    "prompt": "footsteps on gravel, wind",
    "count": 1
  } }

Parameters

Full parameter surface for every model, sourced from gen-ai models info <id> --json. CLI flags show the primary short form; the canonical --kebab-case long form always works too.

mirelo-sfx-2.0-text-to-audio — Mirelo SFX 2.0 (Text to Audio)

Input type: t2a

Param	CLI flag	Type	Values
prompt	-p	text	required (≤2000 chars)
negativePrompt	--neg	text	free text (≤1000 chars)
duration	-d	range	1–600, step 1 (default 30)
count	-n	enum	1 · 2 · 3 · 4 (default 2)
format	--format	enum	wav · mp3 · aac · flac (default wav)
bitrateKbps	--bitrate-kbps	range	32–320, step 1
seed	--seed	range	0–2147483647, step 1

mirelo-sfx-2.0-video-to-audio — Mirelo SFX 2.0 (Video to Audio)

Input type: v2a

Param	CLI flag	Type	Values
videoUrl	--video	file	required video
prompt	-p	text	free text (≤2000 chars)
negativePrompt	--neg	text	free text (≤1000 chars)
preserveSpeech	--preserve-speech	boolean	true · false (default true)
duration	-d	range	1–600, step 1 (default 10)
count	-n	enum	1 · 2 · 3 · 4 (default 2)
format	--format	enum	wav · mp3 · aac · flac (default wav)
bitrateKbps	--bitrate-kbps	range	32–320, step 1
seed	--seed	range	0–2147483647, step 1

mirelo-sfx-2.0-video-to-video — Mirelo SFX 2.0 (Video to Video)

Input type: v2v

Param	CLI flag	Type	Values
videoUrl	--video	file	required video
prompt	-p	text	free text (≤2000 chars)
negativePrompt	--neg	text	free text (≤1000 chars)
preserveSpeech	--preserve-speech	boolean	true · false (default true)
duration	-d	range	1–600, step 1 (default 10)
count	-n	enum	1 · 2 · 3 · 4 (default 2)
seed	--seed	range	0–2147483647, step 1