Hoverfly
Tiny on-device AI models for Kotlin Multiplatform apps
Get picks like this daily. The day's top launches, AI/tech news, and a weekly opportunity spotlight — straight to your inbox.
This is 1 of 64 launches in social media and DM automation tools — see how it stacks up on momentum and crowding →
9 other launches read as similar to this one →
Details
- External ID
- 1266418
- Source
- PH
- Company
- —
- Product
- Hoverfly
- Website domain
- producthunt.com
- Launched
- Oct. 2, 2026
- Cohort
- —
- Upvotes
- 2
- Upvotes percentile
- 0.7630403573137662
- Tags
- Android, Developer Tools, Artificial Intelligence, GitHub
- Fetched at
- Oct. 4, 2026, 1:01 a.m.
- Updated at
- Oct. 4, 2026, 1:01 a.m.
Description
Hoverfly is 8 tiny AI models you add to a Kotlin Multiplatform app: emoji suggestions, smart replies, language detection (Hinglish too), punctuation, toxicity detection, personal-info hiding (phone, UPI, Aadhaar), emotion detection and doodle recognition. They run fully on the device (Android, iOS, desktop, web) in milliseconds, 2–9 MB each. No server, no API key, nothing leaves the phone. Pure Kotlin, on Maven Central, free up to 10,000 monthly active devices.
Enrichment
- Niche
- social media and DM automation tools
- Vertical
- Horizontal
- Function
- Model & infra
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Commercial product
- Normalized one-liner
- on-device ai models for kotlin multiplatform developers
- Manually corrected
- False
Could you build this?
No Training, distilling, and quantizing eight specialized lightweight AI models for cross-platform on-device execution requires deep machine learning research and edge deployment expertise.
What it would actually take: The stack requires a PyTorch training pipeline with curated datasets for Hinglish NLP, Indian PII (Aadhaar, UPI), toxicity, and sketch recognition. Models must be pruned and quantized (int8/FP16) down to single-digit megabytes and exported to ONNX/TFLite/CoreML formats. Finally, low-level Kotlin Multiplatform bindings must interface with platform-specific hardware accelerators (Apple Neural Engine and Android NNAPI).
Competitors
Other products that read as similar to this one — 9 launches clear the similarity bar, closest 8 shown.
Attention rank: #2 of 10 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 297 days after the earliest competitor.
- DeChecker · hn · 2025-12-09 · 6 upvotes · similarity 0.34
- ZeeroGPT · ph · 2026-09-25 · 1 upvotes · similarity 0.32
- OmniChat AI PRO · ph · 2026-09-11 · 1 upvotes · similarity 0.32
- Kodwala · ph · 2026-09-12 · 3 upvotes · similarity 0.32
- AYN KOL AI · ph · 2026-09-13 · 1 upvotes · similarity 0.31
- KoruTalk · ph · 2026-09-28 · 1 upvotes · similarity 0.31
- Huxly · ph · 2026-09-28 · 1 upvotes · similarity 0.31
- Crecaly · ph · 2026-09-21 · 1 upvotes · similarity 0.31
Other launches for this product
Same idea, different domain
Nobody's really built a model & infra tool for Fintech yet.