A practical path through QINGYAN: first run, importing an old draft, reviewing AI proposals, managing credits, and keeping a recoverable copy of your work.
Use the sample book to understand the workflow before bringing in a full manuscript.
Keep the original as a backup, then import one volume or arc at a time.
Treat each generation as a defined writing task rather than an endless chat.
Do not leave the only copy of a manuscript in one device or one tool.
Outlines, character files, foreshadow states, and existing prose are compiled into structured context and injected per task. Output returns as proposals, and only after continuity checks and your line-by-line review does it merge — the structure holds, so the memory holds.
In the current version, work files are stored on your device by default. Sign-in, subscriptions, credits, and AI tasks are handled by QINGYAN services. When you start an AI task, the context needed for that task is sent to the selected service; review the scope before submitting.
Yes. Outline editing, character files, drafting, search, and export are all available offline; reconnect anytime to summon the AI assistant.
No. AI output appears as a proposal or diff first. It is written into the work only after continuity checks and your line-by-line review; you can reject a proposal when it does not fit.
One account can be signed in on up to 3 devices at once. In the current version, each device keeps its own work library, while subscriptions and credits are shared across devices.