Choosing an on-device speech-to-text model: FunASR, Whisper, Vosk, Xiaomi MiMo, and the open ASR field
A practical comparison of open speech-to-text models by streaming support, language coverage, model size, and mobile deployment fit.
Story archive for WayDigital in this collection.
45 public articles
A practical comparison of open speech-to-text models by streaming support, language coverage, model size, and mobile deployment fit.
从模型大小、流式能力、语种覆盖和手机端可行性出发,比较 FunASR、sherpa、Vosk、Whisper、小米 MiMo、PaddleSpeech、WeNet、MiniMind/MiniCPM-o 等开源 ASR 方案。
An updated practical guide to Hermes Agent profiles, now with X-sourced context on multi-profile coordination, Kanban, xurl/browser skills, and Agent OS discussions.
一篇更新后的 Hermes Agent profile 指南:补入 X 上关于多 profile 协调、Kanban、xurl/browser skills 和 Agent OS 讨论的公开线索。
一篇基于官方文档、源码与 PR 信息的中文解读:这个可选运行时解决什么问题、如何启用、适合哪些开发场景。
A practical, source-grounded guide to Hermes Agent’s optional Codex app-server runtime and how it changes coding workflows.
A source-grounded comparison of OpenHuman, OpenClaw, and Hermes Agent across product logic, memory architecture, auto-fetch risks, and possible integration paths.
基于公开仓库、官方文档与代码,对 OpenHuman、OpenClaw 和 Hermes Agent 的产品逻辑、记忆架构、20 分钟自动拉取风险与可结合方向做事实分析。
基于官方仓库、源码和公开使用信号,对 9Router 与 New API 做一次事实对比。
A factual comparison of 9Router and New API: coding-tool router versus multi-user AI gateway.
A practical walkthrough of Hermes Agent Kanban for app development, using a HabitLoop app example to explain boards, tasks, dependencies, dispatcher behavior, worker handoffs, retries, and review loops.
以 HabitLoop 习惯追踪 App 为例,说明 Hermes Agent Kanban 如何用 board、task、link、dispatcher、worker tools 串起产品、设计、工程、QA 和发布流程。