FAQ
Frequently Asked Questions
The questions we hear most about NovelThread. Not answered here? Reach out via the Contact page.
How is NovelThread different from ChatGPT or other chat tools?
NovelThread organizes your manuscript alongside your cast, rules and planted clues, giving AI suggestions relevant story context. This can help with consistency, but AI can still invent details or miss information. Check its suggestions against your manuscript.
Does it work offline without an account?
Yes, completely. NovelThread is a local-first desktop app. Your projects live in a local database on your hard drive, allowing you to write, plan chapters, and manage your story bible with no account and zero internet connection. Network is only needed for cloud models and cross-device sync.
Where is my manuscript stored? Is it uploaded?
Your manuscript stays on your computer by default. It leaves your machine in only two cases: you subscribe and explicitly enable cloud sync (off by default); or you invoke an AI feature, in which case only the relevant passages needed for that request are sent to the model provider.
Will NovelThread train models on my writing?
NovelThread does not use your creative writing to train models or sell user data. The desktop app has no advertising or behavioral analytics. External model and tool providers, including your chosen BYOK provider, process submitted material under their own policies. Review those policies before sending content.
Can the AI change my manuscript without asking?
Never unprompted. Inline rewrites show previews and only replace text when you click confirm. In the assistant, edits require your explicit approval unless you choose automatic mode. A version snapshot is saved before each change, so you can roll back anytime.
Which AI models are supported? Can I bring my own API key?
Yes. NovelThread is not tied to one model: configure any OpenAI-compatible provider with your own API key for a direct device-to-provider connection, or sign in and use the official cloud models with no extra setup.
Which official models are in the catalog?
There are 12 text models available: Kimi K3, GLM 5.3, GLM 5.3 Flash, Gemini 3.8 Flash, DeepSeek V4.1 Flash, Qwen3.8 Flash, MiMo-V2.6-Flash, Hy3, GPT-6 Luna, GPT-6 Sol, Claude Opus 5.5, Claude Sonnet 5.5 Free, Plus, Pro and Max all include every official model.
How do built-in models compare with using my own API key?
Built-in models need no extra setup and use NovelThread AI credits. An active Plus subscription or higher also lets you bring your own API key (BYOK): requests go straight from your device to your chosen provider, which bills you separately. This does not use NovelThread credits. Your key stays in local secure storage; available features depend on the provider.
Is the free tier enough? What does paying add?
The free tier includes local writing and exports, forever. Plus and higher subscriptions add support for your own API key, larger official AI quotas and cross-device sync. Using your own API key does not consume NovelThread credits, but your model provider bills you separately.
How do I cancel? Can I get a refund?
Cancel any time in Settings → Account; access continues to the end of the current billing period. Once charged, fees are not voluntarily refundable or prorated, except where required by law or for cases such as duplicate or unauthorised charges. See the Terms of Service.
Which platforms are supported?
NovelThread currently supports macOS 12 Monterey or later on Apple Silicon. A Windows 10 or later (x64) release is in active preparation and coming soon. Optional voice dictation requires macOS 26 or later; every other writing feature follows the macOS 12 baseline.
Can I import an existing manuscript? Can I export?
Yes. Import from .txt / .md with automatic chapter splitting; export to plain text, Markdown, EPUB or Word. Both run entirely on your machine — nothing is uploaded.