Tag
OpenAI宣布Decisions API正式可用，基于GPT-6 Luna，面向有界答案的快速决策场景，速度比通过Responses API调用GPT-6 Luna快达10倍，并新增GPT-Live原生集成、提示缓存与推理驻留等功能。
Koi 是一款 Mac 应用，将听写工具、会议记录器和 AI 助手三合一，支持离线无限听写、屏幕感知的 AI 助手以及无需机器人入会的通话录制与笔记生成。
Photon 2.0 发布：开发者可以用一键部署将 AI agent 同时带到 iMessage、语音、邮件、WhatsApp、Telegram 和短信平台，推动消息类 agent 走向主流。目前处于公测阶段，并为开发者提供 $100 免费额度。
OpenAI 宣布推出 GPT-Live-1.1，为 ChatGPT、Codex 等产品带来新声音、更好的口音表现以及更少的延迟，并新增法语、德语、印地语、墨西哥西班牙语、韩语和日语语音，今日向所有用户推送，API 支持即将到来。
LlamaForge v0.16 推出了一款免费开源的 llama.cpp 浏览器控制面板，新增了名为 Embers 的本地代理功能，要求代理逐字引用来源并由代码校验引用真实性，另附本地语音播报功能。
Dodaios 的维护者展示了一款 Apache 2.0 开源桌面 AI 代理应用，可在本地 Ollama 模型（Qwen3.5:9b 和 Qwen3.8-16gb）上运行多代理任务，并坦诚说明了本地模型在工具调用上的局限。
Marv is an AI cursor companion that lets users hold a hotkey and ask about anything on screen, answering out loud while drawing arrows, circling buttons, and annotating the display to guide tasks like debugging code, writing spreadsheet formulas, or navigating apps.
Google has released new Gemini models, including Gemini 3.8 Live and 3.5 Transcribe, to enhance real-time voice applications with improved conversational AI and multilingual transcription for developers.
TechCrunch's Equity podcast interviews Sandbar co-founder and CEO Mina Fahmi about the Stream smart ring and why voice is the future of AI wearables.
Airy is a free, fast, and simple voice content creation tool showcased on Hacker News.
The Gemini Live API enables low-latency, real-time voice and vision interactions with Gemini, supporting continuous streams of audio, images, and text for building natural conversational agents.
Heard is a product that gives Claude Code and Codex a voice, enabling voice interaction with these AI coding assistants.
ChatGPT Voice is now available on desktop for all paid plans, enabling voice interaction with the AI assistant.
OpenAI has launched ChatGPT voice functionality on desktop, supporting voice control of the computer and coordinating multiple agents, based on GPT-Live, available worldwide starting today.
OpenAI has launched ChatGPT Voice in the desktop app, allowing users to control their computer and direct multiple agents using voice, powered by GPT-Live. Rolling out globally on macOS and Windows to paid plans.
OpenCode Superapp combines the power of Codex with locally hosted AI models and voice capabilities for coding.
The article discusses the significance of voice technology in artificial intelligence.
A private voice ring for dictation across iOS and Mac, enabling universal dictation via Stream.
The Cicero project enables self-hosted, bidirectional voice conversation with AI agents, supporting local TTS/STT and integration with various agent protocols and tools like Claude Code and Hermes Agent.