Messages that involve conflict, disagreement, or self-advocacy (like correcting someone, setting a boundary, pushing back, or asking for what you need) tend to trigger strong emotional and social pressure. This shows up across emails, texts and phone calls, and takes place in many personal and professional situations.

The messages often sit in drafts for predictable reasons — fear of the reaction, conflict aversion, uncertainty about wording, power dynamics, guilt, and the fact that text strips away vocal cues so a message can read harsher than intended.

There are many contexts and environments where someone face difficulty communicating their needs, defending themselves and voicing their concerns.

Some of the most common situations involve speaking up in close relationships, authorities, and close-knit structures.

When using vibe-coding in my process, I believe the quality of AI generated design depends on how careful, detailed and thorough the PRD is prepared.

To make sure Claude focus on the right problem in the right scope, I translated my initial research and problem framing into explicit descriptions of the problem space and handpicked the most critical aspects of each category of users and scenarios.

To achieve coherent user flows, I outlined 3 user journeys for each of the main categories. For design quality, to ensure the generated output addresses the key functions I have envisioned, I translated my own initial wireframes into detailed, feature-oriented descriptions of the expected user interaction and software behaviors.

As a 0→1 product, Betterwords is set out to have a high bar for consumer-facing design quality from the start.

I first defined the design language and visual communication preferences, then generated UI components using Claude Design. This approach allows me to ensure a cohesive and consistent user experience from the start.

The Betterwords.ai landing page greets the user with a calm, paper-and-ink aesthetic. It promises on a supportive writing companion: it helps you write the messages that matter most and shows you how each way of saying it is likely to land.

A great message starts with understanding the situation. Betterwords.ai clarifies user's intent, context and concerns through a few targeted questions first.

Based on the initial input, Betterwords returns a few drafts with distinct strategies.

For each draft, Betterwords.ai help user visualize the risk level and a likely reaction, so the trade-off is explicit.

User can select any sub-text and tell Betterwords in which direction they want to wording to change. With a just little hint, Betterwords will fill the blanks and generate a story that blends nicely into the context.

To add new content or perspective, user can pick where in the draft they want to elaborate, and tell Betterwords what subject to generate.

Based on the context, user can let Betterwords come up with what additional content and perspectives can be added to the draft.

Betterwords also allows users to tune the message until it sounds like them. The "Tone" and "Length" sliders reshape the draft live, and the "Pros" and "Cons" auto-updates as user tune their message.

Lastly, the "why this works" panel and the likely reaction stay in view as the user edits, so they never lose the thread of what's the pros and cons of this version.