What Shandianshuo is
Shandianshuo (Chinese name 闪电说) is a communication assistant developed by 探未（武汉）科技有限公司 (Tanwei Technology, Wuhan). The official site positions it as a “communication agent” whose goal is to help you communicate better — whether with friends, colleagues, clients, or AI. It combines voice input, context awareness, memory, and task execution around everyday conversations. The entry point is shandianshuo.cn.
Features and boundaries
The official site groups its core capabilities as follows:
- Direct speech (直接说)： press a shortcut, speak, and the cleaned-up result lands in any input box such as WeChat or Feishu. The official wording stresses that the output is “not a verbatim transcript” — it is structured, de-duplicated, corrected, and can use a personal dictionary for names and terminology.
- Drafted replies (帮我说)： hold the shortcut and the AI drafts a reply using on-screen context and stored memory, aimed at situations where you would rather not write it yourself.
- Recall (全面回忆)： before replying, it loads memory about the other person — work preferences, expression habits, past context — and keeps updating it after conversations.
- Knowledge base and skills (技能)： the knowledge base holds style, dictionary, and canned-reply assets; a “skill” can run code, request data, or call tools before the reply is composed.
As a product category, Shandianshuo is a standalone desktop communication tool rather than an input method: it is triggered by shortcuts and works across WeChat, WeCom, DingTalk, Feishu, Qianniu, WhatsApp, browsers, and email. The site offers Windows and macOS downloads and marks the Android, iOS, HarmonyOS, and Linux versions as in internal beta; current availability belongs to the official site. A pricing entry exists, but no specific prices are listed publicly.
Discussion in the show
In Weekly #003’s chapter “语音输入产品与实时反馈” (the voice-input chapter of the Chinese transcript), while discussing how Tencent’s Chatterfly attaches skills to voice input, Xiangyang Qiaomu says “Shandianshuo did this very early”. The hosts then weighed its trade-offs: Guizang points out that its skills are essentially prompts — they cannot operate the computer or execute tools; Xiangyang Qiaomu and Yang Pan felt the product had been made too complex, and Orange adds that the trend now is toward simplicity, with prompt templates everywhere being simplified. These are the participants’ comments in a voice-input context, not official statements. The same discussion also covers Typeless and similar voice-input products.
Frequently asked questions
Where is the Shandianshuo official website?
The official site is shandianshuo.cn, with product information, downloads, and a pricing entry.
Which platforms does Shandianshuo support?
As of September 2026, the official site offers Windows and macOS downloads and marks the Android, iOS, HarmonyOS, and Linux versions as in internal beta. Current availability belongs to the official site.
How is Shandianshuo different from a voice input method?
A voice input method types what you say; Shandianshuo’s “direct speech” restructures and corrects what you say, and on top of that it drafts replies from context, keeps communication memory, and runs skills that execute tasks. It is a desktop tool covering the whole conversation flow, not an input method.
How much does Shandianshuo cost?
The site has a pricing entry but does not list specific prices. Whether there is a free tier and what each plan includes should be checked on the official site.