Analysis

Speech synthesis does not have to visit a server. Gradium says its Phonon models run locally, without a network for inference.

The November 5 post explains an earlier release. Phonon 31m English is Apache 2.0; the 40m multilingual model has a non-commercial licence, so the editions are not interchangeable for commercial use.

Local speech moves network, data-handling and hardware questions to the product team. Gradium's performance claims are the authors' own, so pronunciation, speed and licence terms still need checking on intended hardware.

Source: A Closer Look at Phonon TTS. Launching High-Quality On-Device Speech That Tops the Benchmarks. · 2026-11-05 · Primary source context · attributed account

Why it matters

Local speech shifts deployment and licensing choices onto the product team.

About this article

AI-assisted editorial analysis drafted by Codex as a stand-in using Mira Vale saved persona brief; writer model identity is not exposed. Recorded runtime label: Codex (model identity not exposed); edited by Ellis Finch on Claude claude-sonnet-5-5. Mira Vale is not a human journalist. Silver Owl Journal is responsible for publication decisions and corrections.

Editorial standards & AI production ↗ · About Mira Vale

Sources and editorial note

Codex stand-in draft using Mira Vale saved persona brief on Codex (model identity not exposed) (openai-a); edited and accepted by Ellis Finch on Claude claude-sonnet-5-5. Draft route note: Codex stand-in using the saved Mira Vale editorial brief; no native writer call. Evidence limits: Author-reported capabilities; no independent listening, speed test or licence review.

  • A Closer Look at Phonon TTS. Launching High-Quality On-Device Speech That Tops the Benchmarks. (2026-11-05)