The content-safety layer
for image-heavy products.
Drop in our SDK. Detect explicit content before it reaches your users — on device, in under 50 ms, with confidence scores you can act on.
Swift · iOS
1import AntiNude
2
3let scanner = Scanner(key: "ak_live_…")
4let result = try await scanner.scan(image)
5
6// → result.verdict: .safe
7// → result.scores.nudity: 0.02
Kotlin · Android
1import io.antinude.Scanner
2
3val scanner = Scanner("ak_live_…")
4val result = scanner.scan(bitmap)
5
6// → result.verdict == SAFE
7// → result.scores.nudity == 0.02
↖3 lines · no servers
§01How it runs (on-device)
User image
camera, gallery,
upload
→
AntiNude SDK
CoreML / TFLite
~ 47 ms · ~12 MB
→
Your verdict
.safe / .unsafe
+ scores
↑ image bytes never leave the device · cloud is only used for license check
§02By the numbers
99.94%
precision
on held-out test set
47 ms
median latency
on iPhone 12
100k+
training images
12 cultural contexts
12 MB
SDK size
incl. model weights
§03Compliance & safety
✓ On-device inference
✓ Images never leave the device
✓ No image upload
✓ Encrypted in transit & at rest
✓ GDPR-ready
✓ DPA available