闪电说

Chinese AI voice input for Windows and Mac with a local speech model and an agent mode for drafting replies.

Visit website (opens in a new tab) Compare with other tools

Shandianshuo is a desktop voice input tool from Tanwei (Wuhan) Technology for Windows and macOS. A short press of the hotkey transcribes speech, fixes errors and punctuation and inserts the text into any input box; a long press runs an agent that drafts replies from screen context. It ships a local SenseVoice model and can use cloud or self-hosted LLMs.

Pros

  • Local speech-recognition model and self-configured Ollama or LM Studio endpoints allow offline use at no cost
  • Free tier plus paid plans; users can also bring their own API keys
  • Custom dictionary, adjustable polish level, spoken edit commands on selected text and local history

Cons

  • Interface and documentation are Chinese-only; no published language count
  • Free plan limited to 0.5 hours of dictation per month on the hosted models
  • Mobile and Linux versions are still in closed beta
Platforms
  • macOS
  • Windows
Speech processing
Hybrid
Works offline
Yes
Languages
—

Pricing

Pricing model
Freemium
Free tier
0.5 h dictation + 50 agent credits per month
Pro (monthly) ¥29/mo
Pro (annual, ¥19.9/mo) ¥238.80/yr
Max (monthly) ¥139/mo
Max (annual, ¥99/mo) ¥1188/yr

Features

  • Yes Works in any app
  • No Real-time transcription
  • Yes Auto punctuation
  • Yes AI cleanup
  • Yes Custom vocabulary
  • Yes Voice commands
  • Yes Push to talk
  • No Auto language detection
  • No Translation
  • No File transcription
  • No Meeting notes
  • No Multilingual interface
  • Yes History

Compare 闪电说 with

Last verified 2026-09-19