algolia/voice-overlay-ios

What it solves

This library provides a polished, ready-to-use user interface overlay for iOS apps to convert speech to text. It simplifies the process of implementing voice input by handling the complex parts of the user experience, such as requesting necessary microphone and speech recognition permissions and guiding users to settings if permissions are denied.

How it works

The project uses Apple's native SFSpeechRecognizer to perform the actual speech-to-text conversion. It wraps this functionality in a VoiceOverlayController that manages the recording state, provides visual feedback (like ripples), and returns the transcribed text via a handler or a delegate.

Who it’s for

iOS developers who want to add voice search or voice input capabilities to their applications without building the custom UI and permission-handling logic from scratch.

Highlights

• Permission Management: Automatically handles microphone and speech recognition requests and guides users to system settings if permissions are missing.
• Customizable UI: Allows developers to modify colors, titles, and layout constants for different screens (e.g., permission screen, input screen).
• Flexible Recording: Supports auto-start and auto-stop (based on silence timeouts) or manual trigger modes.
• Custom Speech Controller: Developers can implement the Recordable protocol to use their own custom speech recognition logic.
• Result Screen: Includes a beta feature to show a processed result screen after recording is finished.