App · iPhone
NEIROA
On-device AI music generation for iPhone. Tracks up to 120 seconds. 10 successful generations free. No account required.
English/日本語
Powered by Stability AI
Overview
On-device AI music generation for iPhone.
NEIROA is a MONOFRAME app that creates music on your iPhone from a prompt or a simple set of choices. Tracks can be up to 120 seconds. No account is required. Finished tracks live in a local library, and you can export them as WAV files.
Pricing
10 successful generations free. No subscription.
The first 10 successful generations are free. Failed generations do not count. After that, a one-time purchase of NEIROA Unlimited removes the generation limit. There is no subscription. Prices vary by country and region. In Japan, NEIROA Unlimited is ¥1,980.
On Device
Prompts and audio do not leave the phone for generation.
Music generation uses Stable Audio 3 Small Music on the device. Prompt AI can also run on the device when that model is installed. NEIROA does not send prompts or generated music to MONOFRAME servers. Songs, playlists, and related data stay on the iPhone unless you export or share them.
Availability
Available on the App Store.
Built for iPhone. After the models required for music generation are on the device, generation and playback can run offline. Prompt AI is only for turning everyday language into a music prompt; music generation still works if it is not installed. Downloading models, completing a purchase, and restoring a purchase may require the internet.
Performance Test
A high-temperature stress test. These times are not typical single-run speed.
On iPhone 17 with NEIROA v1.0 App Store, 30 consecutive 120-second tracks were generated. That is 60 minutes of music. Total generation time was 11:24. The overall average was 22.8 seconds per track. The first 5 tracks took 95.16 seconds (19.03 seconds per track). The last 5 tracks took 130.36 seconds (26.07 seconds per track). A separate single-run measurement reached about 11 seconds. Results vary with environment and device state.
Conditions
- Case on
- Connected to a Mac over USB for power
- Thermal State = Serious at the start
- Timed in Instruments
- GPU load was not measured
Time by 5-track group
| Tracks | Time |
|---|---|
| 1–5 | 95.16 seconds |
| 6–10 | 104.16 seconds |
| 11–15 | 111.57 seconds |
| 16–20 | 119.10 seconds |
| 21–25 | 123.65 seconds |
| 26–30 | 130.36 seconds |
Reference measurement
Separate cooled single-run measurements of 120-second generation. Not part of the 30-track stress test.
These are reference values. Measurement conditions differed, so this is not a strict comparison between devices.
| iPhone 17 | about 11 seconds |
| iPhone 15 Plus | about 21 seconds |
Create
Use Easy or Detailed creation, then generate. Prompt AI can turn everyday language into a music prompt when that model is installed.
Library
Completed tracks are stored on the device. Share or export a WAV from the track menu, and keep important pieces off-device as well.
Privacy
No account, no ads, and no MONOFRAME analytics inside the app. Photo access is used only when you choose a playlist image.