Core criteria
- Workflow fit: which writing stage, project type, and level of technical comfort the product serves.
- Project context: how the system organizes manuscript, outline, research, and story facts for the writer and any AI features.
- Review and recovery: whether changes are visible, versioned, and reasonably reversible.
- Ownership and portability: where the canonical work lives and how completely it can leave the product.
- AI control: how the writer chooses context, models, scope, and the acceptance of generated changes.
- Maturity: whether the capability is released, documented, stable enough to evaluate, or merely planned.
How products enter a list
We begin with the reader’s job, not a vendor category. A product must be meaningfully usable for that job and have enough first-party documentation to support accurate claims. We link to official documentation and record the article’s verification date.
How we use “best”
Each selection is named for a fit: best for guided fiction generation, best for local Markdown control, best for publication-oriented export, and so forth. Our overall conclusion may be that no single product dominates every criterion.
Unreleased products
An unreleased product cannot win a current recommendation. It may appear in a clearly labeled “watch” section if its stated direction helps readers understand an emerging workflow. Plans are not scored as shipped features.
Commercial influence
No placement can be purchased. We currently use no affiliate links. If that changes, compensation will be disclosed next to affected links and will not alter eligibility criteria.
Corrections and updates
Material corrections are recorded on the corrections page. Product changes are reflected in the updated date and, when consequential, an editor’s note.