speech-recognition

Name: speech-recognition
Author: dpearson2699

2.5kinstalls45forks

How to install speech-recognition?

dpearson2699/swift-ios-skills/speech-recognition

$npx -y skills add dpearson2699/swift-ios-skills --skill speech-recognition

Installs into the current project.

Summary

Transcribe speech to text using Apple's Speech framework. Use when implementing live microphone transcription with AVAudioEngine, recognizing recorded audio files, handling speech and microphone authorization, choosing on-device vs server-backed SFSpeechRecognizer behavior, or ad

Files · 1

View on GitHub

SKILL.md

1	# Speech Recognition
2
3	Transcribe live and pre-recorded audio to text using Apple's Speech framework.
4	Covers `SpeechAnalyzer` / `SpeechTranscriber` (iOS 26+) and
5	`SFSpeechRecognizer` (iOS 10+) fallback guidance.
6
7	Scope boundary: Use this skill for speech-to-text recognition, speech
8	authorization, microphone capture plumbing, and result handling. Hand off text
9	analysis, language identification after transcription, sentiment, embeddings,
10	and translation to `natural-language`; hand off audio playback UI to `avkit`;
11	hand off summarization or generation over transcripts to `apple-on-device-ai`.
12
13	## Contents
14
15	- [SpeechAnalyzer Strategy (iOS 26+)](#speechanalyzer-strategy-ios-26)
16	- [SFSpeechRecognizer Setup](#sfspeechrecognizer-setup)
17	- [Authorization](#authorization)
18	- [Live Microphone Transcription](#live-microphone-transcription)
19	- [Pre-Recorded Audio File Recognition](#pre-recorded-audio-file-recognition)
20	- [On-Device vs Server Recognition](#on-device-vs-server-recognition)
21	- [Handling Results](#handling-results)
22	- [Common Mistakes](#common-mistakes)
23	- [Review Checklist](#review-checklist)
24	- [References](#references)
25
26	## SpeechAnalyzer Strategy (iOS 26+)
27
28	Use `SpeechAnalyzer` for modern iOS 26+ speech analysis, especially long-form
29	recordings, live transcription, time-indexed transcripts, and fully on-device
30	flows. Keep `SFSpeechRecognizer` for iOS 10+ deployment targets, server-backed
31	locale coverage, or existing callback/delegate implementations.
32
33	Read [SpeechAnalyzer patterns](references/speechanalyzer-patterns.md) when
34	implementing an iOS 26+ transcription pipeline, model asset handling, volatile
35	results, or file/buffer examples.
36
37	### SpeechAnalyzer setup checklist
38
39	1. Choose the module:
40	- `SpeechTranscriber` for the newer general-purpose on-device model.
41	- `DictationTranscriber` when `SpeechTranscriber` is unavailable for the
42	current device or locale and dictation-compatible support is acceptable.
43	- `SpeechDetector` only in conjunction with a transcriber when voice
44	activity detection is worth the accuracy/power tradeoff.
45	2. Check support before creating the session:
46	- `SpeechTranscriber.isAvailable`
47	- `SpeechTranscriber.supportedLocale(equivalentTo:)`
48	- `SpeechTranscriber.installedLocales` / `supportedLocales` when showing
49	language choices.
50	3. Pick a documented preset:
51	- `.transcription` for basic accurate transcription.
52	- `.progressiveTranscription` for live UI updates.
53	- `.timeIndexedProgressiveTranscription` when playback highlighting needs
54	`audioTimeRange`.
55	4. Install required assets with `AssetInventory.assetInstallationRequest`.
56	5. Convert live audio buffers to
57	`SpeechAnalyzer.bestAvailableAudioFormat(compatibleWith:)` before yielding
58	`AnalyzerInput`.
59	6. Consume module results from their `AsyncSequence` in a separate task.
60	7. Finish explicitly with `finalizeAndFinish(through:)`,
61	`finalizeAndFinishThroughEndOfInput()`, or `cancelAndFinishNow()`.
62
63	Do not use an `offlineTranscription` preset; Apple does not document one.
64	Finishing an `AsyncStream` input sequence does not finish the analyzer session.
65
66	## SFSpeechRecognizer Setup
67
68	### Creating a recognizer with locale
69
70	```swift
71	import Speech
72
73	// Default locale (user's current language)
74	let recognizer = SFSpeechRecognizer()
75
76	// Specific locale
77	let recognizer = SFSpeechRecognizer(locale: Locale(identifier: "en-US"))
78
79	// Check if recognition is available for this locale
80	guard let recognizer, recognizer.isAvailable else {
81	print("Speech recognition not available")
82	return
83	}
84	```
85
86	### Monitoring availability changes
87
88	```swift
89	final class SpeechManager: NSObject, SFSpeechRecognizerDelegate {
90	private let recognizer = SFSpeechRecognizer()!
91
92	override init() {
93	super.init()
94	recognizer.delegate = self
95	}
96
97	func speechRecognizer(
98	_ speechRecognizer: SFSpeechRecognizer,
99	availabilityDidChange available: Bool
100	) {
101	// Update UI — disable record button when unavailable
102	}
103	}
104	```
105
106	## Authorization
107
108	Request both speech recognition and microphone permissions before starting
109	live transcription. Add these keys to `Info.plist`:
110
111	- `NSSpeechRecognitionUsageDescription`
112	- `NSMicrophoneUsageDescription`
113
114	```swift
115	import Speech
116	import AVFoundation
117
118	func requestPermissions() async -> Bool {
119	let speechStatus = await withCheckedContinuation { continuation in
120	SFSpeechRecognizer.requestAuthorization { status in
121	continuation.resume(returning: status)
122	}
123	}
124	guard speechStatus == .authorize

People who liked this also viewed

6 picks

Type

speech-recognition

2.5kinstalls45forks

View on GitHub

How to install speech-recognition?

dpearson2699/swift-ios-skills/speech-recognition

$npx -y skills add dpearson2699/swift-ios-skills --skill speech-recognition

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Speech Recognition
2
3	Transcribe live and pre-recorded audio to text using Apple's Speech framework.
4	Covers `SpeechAnalyzer` / `SpeechTranscriber` (iOS 26+) and
5	`SFSpeechRecognizer` (iOS 10+) fallback guidance.
6
7	Scope boundary: Use this skill for speech-to-text recognition, speech
8	authorization, microphone capture plumbing, and result handling. Hand off text
9	analysis, language identification after transcription, sentiment, embeddings,
10	and translation to `natural-language`; hand off audio playback UI to `avkit`;
11	hand off summarization or generation over transcripts to `apple-on-device-ai`.
12
13	## Contents
14
15	- [SpeechAnalyzer Strategy (iOS 26+)](#speechanalyzer-strategy-ios-26)
16	- [SFSpeechRecognizer Setup](#sfspeechrecognizer-setup)
17	- [Authorization](#authorization)
18	- [Live Microphone Transcription](#live-microphone-transcription)
19	- [Pre-Recorded Audio File Recognition](#pre-recorded-audio-file-recognition)
20	- [On-Device vs Server Recognition](#on-device-vs-server-recognition)
21	- [Handling Results](#handling-results)
22	- [Common Mistakes](#common-mistakes)
23	- [Review Checklist](#review-checklist)
24	- [References](#references)
25
26	## SpeechAnalyzer Strategy (iOS 26+)
27
28	Use `SpeechAnalyzer` for modern iOS 26+ speech analysis, especially long-form
29	recordings, live transcription, time-indexed transcripts, and fully on-device
30	flows. Keep `SFSpeechRecognizer` for iOS 10+ deployment targets, server-backed
31	locale coverage, or existing callback/delegate implementations.
32
33	Read [SpeechAnalyzer patterns](references/speechanalyzer-patterns.md) when
34	implementing an iOS 26+ transcription pipeline, model asset handling, volatile
35	results, or file/buffer examples.
36
37	### SpeechAnalyzer setup checklist
38
39	1. Choose the module:
40	- `SpeechTranscriber` for the newer general-purpose on-device model.
41	- `DictationTranscriber` when `SpeechTranscriber` is unavailable for the
42	current device or locale and dictation-compatible support is acceptable.
43	- `SpeechDetector` only in conjunction with a transcriber when voice
44	activity detection is worth the accuracy/power tradeoff.
45	2. Check support before creating the session:
46	- `SpeechTranscriber.isAvailable`
47	- `SpeechTranscriber.supportedLocale(equivalentTo:)`
48	- `SpeechTranscriber.installedLocales` / `supportedLocales` when showing
49	language choices.
50	3. Pick a documented preset:
51	- `.transcription` for basic accurate transcription.
52	- `.progressiveTranscription` for live UI updates.
53	- `.timeIndexedProgressiveTranscription` when playback highlighting needs
54	`audioTimeRange`.
55	4. Install required assets with `AssetInventory.assetInstallationRequest`.
56	5. Convert live audio buffers to
57	`SpeechAnalyzer.bestAvailableAudioFormat(compatibleWith:)` before yielding
58	`AnalyzerInput`.
59	6. Consume module results from their `AsyncSequence` in a separate task.
60	7. Finish explicitly with `finalizeAndFinish(through:)`,
61	`finalizeAndFinishThroughEndOfInput()`, or `cancelAndFinishNow()`.
62
63	Do not use an `offlineTranscription` preset; Apple does not document one.
64	Finishing an `AsyncStream` input sequence does not finish the analyzer session.
65
66	## SFSpeechRecognizer Setup
67
68	### Creating a recognizer with locale
69
70	```swift
71	import Speech
72
73	// Default locale (user's current language)
74	let recognizer = SFSpeechRecognizer()
75
76	// Specific locale
77	let recognizer = SFSpeechRecognizer(locale: Locale(identifier: "en-US"))
78
79	// Check if recognition is available for this locale
80	guard let recognizer, recognizer.isAvailable else {
81	print("Speech recognition not available")
82	return
83	}
84	```
85
86	### Monitoring availability changes
87
88	```swift
89	final class SpeechManager: NSObject, SFSpeechRecognizerDelegate {
90	private let recognizer = SFSpeechRecognizer()!
91
92	override init() {
93	super.init()
94	recognizer.delegate = self
95	}
96
97	func speechRecognizer(
98	_ speechRecognizer: SFSpeechRecognizer,
99	availabilityDidChange available: Bool
100	) {
101	// Update UI — disable record button when unavailable
102	}
103	}
104	```
105
106	## Authorization
107
108	Request both speech recognition and microphone permissions before starting
109	live transcription. Add these keys to `Info.plist`:
110
111	- `NSSpeechRecognitionUsageDescription`
112	- `NSMicrophoneUsageDescription`
113
114	```swift
115	import Speech
116	import AVFoundation
117
118	func requestPermissions() async -> Bool {
119	let speechStatus = await withCheckedContinuation { continuation in
120	SFSpeechRecognizer.requestAuthorization { status in
121	continuation.resume(returning: status)
122	}
123	}
124	guard speechStatus == .authorize

People who liked this also viewed

6 picks

Type