SFX 2.0
Our newest model for sound effects from text prompts or video.
Each second of generated audio, per variant
- Studio
Was 15, now 10 credits/sec33% off until December 12, 2026- API
Was 15, now 10 credits/sec33% off until December 12, 2026
Compare how many credits each product uses in Studio and through the Mirelo API.
Rates for products available in Studio, through the API, or both.
| Product | What is measured | Studio | API |
|---|---|---|---|
| SFX 2.0 | Each second of generated audio, per variant | ||
| SFX 1.6 | Each second of generated audio, per variant | 10 credits/sec | 10 credits/sec |
| Preserve Speech | Each second of source audio, once when speech is detected | Not available | +8 credits/sec |
| Music 1.0 | Each second of generated audio, per variant | 10 credits/sec | Not publicly offered |
| Audio-to-MIDI Pro | Each second of source audio | 5 credits/sec | 2.5 credits/sec |
| Lyrics transcription | Each started minute of source audio | +300 credits per started minute | Not publicly offered |
| Voice Isolation | Each second of processed source audio | 2.3 credits/sec | Not available |
Our newest model for sound effects from text prompts or video.
Each second of generated audio, per variant
Generate sound effects from text prompts or video.
Each second of generated audio, per variant
Add-on
Keep detected dialogue while generating sound around it.
Each second of source audio, once when speech is detected
Generate music from a prompt or video.
Each second of generated audio, per variant
Transcribe audio into editable MIDI and notation.
Each second of source audio
Add-on
Time-aligned lyrics alongside an Audio-to-MIDI transcription.
Each started minute of source audio
Remove background sound while retaining speech.
Each second of processed source audio
The SFX 2.0 discount applies to every account automatically. The crossed out rate is what it costs after the discount ends.
Generated-audio rates multiply by the number of variants. Processing rates apply to the source duration once unless a row says otherwise.
Generated and processed audio totals round up to the next whole credit. Audio-to-MIDI rounds to the nearest whole credit, with a one-credit minimum.
+ marks an add-on, charged on top of the product it sits under. Preserve Speech is charged once when speech is detected. Preflight checks that you can cover the possible surcharge; no speech means no surcharge.
Lyrics transcription: Charged even when a track has no sung words.
Not available means Mirelo does not offer that product on the named surface.
Not publicly offered means the product is not available through the public Mirelo API.
These models remain available for compatibility. Use SFX 1.6 for new work.
| Product | What is measured | Studio | API |
|---|---|---|---|
| SFX 1.5 | Each second of generated audio, per variant | 10 credits/secExisting projects | 10 credits/secDeprecated API |
| SFX 1.0 | Each second of generated audio, per variant | 7 credits/secExisting projects | 7 credits/secLegacy API |
Deprecated sound-effects model retained for existing integrations.
Each second of generated audio, per variant
Legacy sound-effects model retained for existing integrations.
Each second of generated audio, per variant
These rates apply to Mirelo Studio and the Mirelo API. Third-party platforms set and bill their own prices.
Compare the monthly plans that include credits for Mirelo Studio and eligible API usage.