Award listed
Original award wording
Apps for Your Life — 1st Place
Project profile
Second Voice
Turns unclear speech into a clearer sentence only after the user confirms it.
Second Voice is an assisted-communication app for people with unclear speech or limited motor control who need to participate in live conversations. The creator describes recording or transcribing speech, checking a personal phrasebook and context, then proposing two or three sentences.
The user selects or edits a sentence before text-to-speech reads it aloud; the system is not meant to invent an unconfirmed statement on the user's behalf. GPT-5.6 is described as handling reconstruction and meaning, with transcription and TTS handling input and output. The creator emphasizes no recorded training corpus, low latency and confirmation, but these are design claims, not tested comparative outcomes.
Creator or team (original public names)
Ravitez Dondeti
Links & sources
Links may have changed since we checked them. Descriptions from creators have not been independently tested.
Details in the original language
Original title
Second Voice
Original creator credit
Ravitez Dondeti
作品是什么
面向构音困难者的语音沟通辅助应用；以下能力来自作者参赛说明，未实际使用。
给谁用
构音困难或运动控制受限、需要在现场对话中被理解的人。
问题或体验
零碎语音不易被理解，既要快速参与对话，也不能让模型未经确认替本人发言。
如何使用
录入语音并转写 → 结合个人短语本和当前情境 → 给出2至3个候选句 → 用户选择或修改 → TTS朗读。
得到什么
可由用户确认的句子和清晰朗读，不是自动替用户决定表达内容。
关键机制
GPT-5.6承担句子重构与歧义处理；转写和TTS分别承担输入/输出，结构化结果与不确定时弃答控制猜测。
特点
作者强调不要求预录训练样本，并保留低操作成本的确认步骤；这是设计主张，非独立效果或竞品比较验证。
Events & award announcements
Read original announcement
Sources last checked: 2026-10-10