The Hugging Face article and two public repositories confirm versions, backends, platform limits, dependency conflict and publisher-run timing.
AnySpeechRecognizer and AnyDecisionModel open Apple-native AI components
2026-10-29Verified announcementTRIAL
AnySpeechRecognizer 0.1.0 and AnyDecisionModel 0.6.0 are open Swift packages for on-device speech and constrained model decisions on Apple platforms. Speech backends cover Parakeet, Sortformer and Apple SpeechAnalyzer; decision backends include MLX, Core AI, llama.cpp, Jev and Ollama. They require Swift 6.2 and current Apple operating systems, and the documented Core AI dependency sets currently conflict when both packages are combined.
What this means for you
For Apple-native product teams, the packages reduce glue code for local speech and typed choices. The narrow platform baseline, dependency conflict and publisher-run M4 Max timing mean teams should test integration and latency before adopting.
Evidence & related material
Technical details & scores
| Versions | AnySpeechRecognizer 0.1.0; AnyDecisionModel 0.6.0 |
|---|---|
| Platform | Swift 6.2+; macOS/iOS 26+; Apple silicon for Parakeet, Sortformer and MLX paths |
| Speech | Parakeet, Sortformer and Apple SpeechAnalyzer; multilingual recognition with punctuation/capitalization where backend supports it |
| Decisions | typed probability, choice and score APIs; MLX, Core AI, llama.cpp, Jev and Ollama backends |
| Known Conflict | AnyDecisionModel Core AI trait uses coreai-kit 0.7.3 with swift-huggingface <1.0, while AnySpeechRecognizer requires 1.x |
| Vendor Timing | M4 Max batch 16 around 0.45 s after load; 343 emoji choices around 10 s |
| Time Basis | Hugging Face feed exposes date 2026-10-29 without time; date is wholly inside the scan window |
Technical impact 3
Engineering value 4
Adoption signal 2
Maturity 2
Verification cost 2
Risk 2