Aquavoice

Aquavoice ships two complementary products from a single speech engine: an AI voice keyboard for iOS and the Avalon speech-to-text API for developers. Avalon is the more technically distinctive of the two, marketed as a drop-in replacement for OpenAI Whisper that requires changing only two lines of code while delivering what the vendor describes as 97.3% accuracy on its proprietary AISpeak benchmark. See the broader AI Speech Synthesis landscape for context on where Avalon fits among transcription engines.

The differentiation claim rests on training data. According to the Avalon API documentation, most speech models are trained on audiobook-quality narration—clean, deliberate, well-enunciated speech. Avalon was instead trained on developer workflows: transcripts paired with CLI sessions and IDE captures. The vendor asserts this yields literal transcript fidelity—preserving the exact command, casing, and model name a developer dictates—rather than the hallucinated model names or misheard commands that can surface with general-purpose ASR models. This is a vendor claim; no independent benchmark validates the AISpeak figure or the literal-fidelity assertion against a standardized test set.

On the consumer side, the iOS app functions as an AI voice keyboard. Downloaded from the App Store, it replaces the system keyboard for dictation use cases, presumably routing speech through the same engine that powers Avalon on the backend. The app is live on iOS at the time of this review; no Android or desktop consumer client appears in the source pack.

Operationally, Aquavoice signals enterprise awareness. A Vanta Trust Center link appears in the site footer alongside a public status page at status.aquavoice.com. A web dashboard at app.aquavoice.com/manage handles account administration. Documentation, a user guide, a blog, and a dedicated FAQ page are linked from the primary navigation. Developer onboarding is straightforward: obtain an API key and integrate the two-line change into an existing Whisper pipeline.

Pricing specifics are absent from the available evidence. The homepage notes that Avalon is included in pro plans and that pro users received priority access to new features during a launch week, but no plan tiers, per-call costs, or free-tier details appear. The multilingual speech model is referenced visually through a globe graphic on the homepage, but no language list or coverage data is provided. Prospective users should verify the multilingual scope and pricing model directly before adoption.