desert-ant-labs/readable-runs
19.7 GB
None defined yet.
A European AI lab building opinionated on-device models and SDKs.
| Model | What it does | Card | Demo | SDKs | Status |
|---|---|---|---|---|---|
| Clear | Audio enhancement, a 10-minute clip in ~7 seconds | Model | Demo | iOS · Android · Web | Available |
| Emo | Multi-language emoji suggestions, as fast as you can type | Model | Demo | iOS · Android · Web | Available |
| Redact | Filter PII from any input, on device, across 24 EU languages | Model | Demo | iOS · Android · Web | Available |
| Shapes | Anyone can draw, with fast cross-platform shape matching | Model | Demo | iOS · Android · Web | Available |
| Uhm | SOTA filler-word detection, zero inference cost | Model | Demo | iOS | Available |
| Eye | Pick the best images or frames. | Closed Beta | |||
| Face | Lightning-fast face matching and search in your app | Closed Beta | |||
| Moderator | On-device nudity detection before upload | Model | Closed Beta | ||
| Schemer | Free text to typed JSON, no hallucinated fields | Model | Demo | Closed Beta | |
| Toxic | Hate-speech and abuse triage in eight languages, fully on device | Model | Demo | Closed Beta | |
| Who | Fast, on-device audio visual speaker diarization. | Closed Beta |
Every model is free up to 100k Monthly Active Devices. Unlimited inference per device.
Drop-in SDKs for iOS, Android, and Web ship with each available model, linked in the table above. More platforms and models are on the way.
Email contact@desertant.ai for early access.
On-device language ID — a 2 MB model that runs in-browser
Multilingual hate-speech triage scores, fully on device
Every uh, um, hmm tagged at 20 ms, on device, multilingual
Free text in, typed JSON out. Your schema, filled.
On-device multilingual PII redaction, in the browser.