Show HN: Revise – An AI Editor for Documents
HCI Today summarized the key points
- •This article introduces a web-based document editor that lets users choose from the latest AI models by OpenAI, Anthropic, and xAI.
- •The editor is a new Canvas-rendering-based document editor that allows direct selection of multiple AI models.
- •In the discussion, the developer was highly praised for the polish achieved through 10 months of solo work and for tackling a challenging domain.
- •At the same time, specific bugs and usability issues were pointed out in detail features like lists, selection, shortcuts, and table editing.
- •Overall, the potential and polish were well-regarded, but finer editing improvements are still needed for it to work as a practical writing tool.
This summary was generated by an AI editor based on HCI expert perspectives.
Why Read This from an HCI Perspective
This article shows what becomes decisive for user experience (UX) when AI writing tools evolve from simple sentence generators to actual editing environments. Details like model selection, document editor architecture, shortcuts, selection ranges, and formatting preservation directly determine usability quality. For HCI/UX practitioners and researchers, it is a worthwhile case study of how generative AI is reshaping traditional document workflows.
CIT's Commentary
From CIT's perspective, this product is not just 'a document editor with AI bolted on,' but an experiment in reducing the interaction cost of document creation. The recurring feedback in top comments emphasizes that low-level interaction quality—focus persistence, selection consistency, platform convention adherence—matters more than the presence of features. Even with generative AI, users still judge by the predictability of everyday operations like cursor movement, selection, undo, and formatting propagation. The demand to 'not look like AI' or 'help write more clearly' suggests that reducing cognitive load may be the core value, beyond model performance. CIT believes evaluating such tools should consider not just output quality, but also friction in the writing flow, alignment with platform expectations, and how well it supports editorial judgment.
Questions to Consider While Reading
- Q.What editing interaction qualities must this editor guarantee before AI assistance to earn user trust?
- Q.If adding features for logic and clarity rather than sentence generation, what feedback method would be most useful?
- Q.How strictly should platform-specific shortcuts and selection behaviors be followed in a generative editor?
This commentary was generated by an AI editor based on HCI expert perspectives.
Please refer to the original for accurate details.
Subscribe to Newsletter
Get the weekly HCI highlights delivered to your inbox every Friday.