The beta build of WhatsApp for iOS has introduced a feature that can convert voice messages into written text, according to observations from the WABetaInfo portal. The news initially surfaced around the mid-2021 period, with early indications that this capability would eventually appear in the app. Subsequent investigations into the feature stalled for reasons that were not publicly disclosed, leaving questions about timelines and readiness.
Recent testing of WhatsApp version 23.3.0.73 on iOS reveals a functioning transcription option accompanied by a dedicated in-app notification. The message notes outline several conditions that can lead to transcription errors, such as parts of speech that are not captured accurately or instances where the transcription language does not align with the device system language. These notes help users understand why some voice messages may not be transcribed perfectly and what adjustments might improve results.
Importantly, the transcripts appear to be produced and decrypted on the user’s device, ensuring that only the participants in the chat can read the converted text. This on-device processing approach aims to preserve privacy by preventing remote servers from accessing the spoken content of conversations.
The transcription feature is currently still listed as a testing capability, with no official word on when it will roll out to stable versions across platforms such as Android and iOS, or when an expanded feature set might be expected. The absence of a firm release window means users should anticipate ongoing refinement and potential changes before a wider launch occurs.
In the broader messaging landscape, a comparable option exists in Telegram’s platform, though it is restricted to users holding a paid subscription. The expectation for WhatsApp is that transcription capabilities could eventually become available to all users without similar access barriers once a stable version is released and validated across environments.
Earlier discussions in tech coverage suggested additional broadcast-oriented audio features could accompany voice messaging, though current official confirmations focus primarily on transcription. The evolving capabilities reflect a desire to provide faster, text-based scannability of spoken content while maintaining end-to-end privacy safeguards and platform-wide compatibility improvements. Marked notes from ongoing testing cite practical limitations and language-compatibility considerations as central factors guiding the final design and rollout decisions.