What does the AI chat shell for a SaaS product prompt produce?

A saas website with Sidebar, Message thread, Composer, Model chip. The vibe target is "focused, calm, productive" and the prompt body is printed in full on the page.

How well does it score?

91% average fidelity over 3 recorded runs. Best model: Claude 4.6 Sonnet. Weakest recorded run: 89% on GLM-4.6 (CN).

Are failures included?

No build failures are recorded for this prompt, and the log says so rather than implying every model succeeded elsewhere.

Concept render — composed live from this prompt's palette & layout. Reproducible with the library, not a stock screenshot.

Concept scene: Star Motes — open in library →

SaaSFeaturedby Motif Studiopublished 2026-09-14

AI chat shell for a SaaS product

zh-CN: SaaS 产品的 AI 聊天外壳

vibe: focused, calm, productive

starter — 4 blocks, 2 stacksany viewport≈ 320 lines generated (estimate)≈ 600 tokens to run (estimate)

prompt.txt — copy-paste ready

Paste into:

This is the text every run below was produced from. The other targets wrap it for a specific tool — same brief, different framing.

A full AI chat shell for a SaaS product: a collapsed sidebar listing past threads, a message thread where assistant messages carry a model chip and user messages are right-aligned, a composer with Enter-to-send and Shift+Enter for a newline, and a "thinking" state shown as breathing dots before each answer. Calm neutral palette, one accent for the assistant, no fake avatars. The composer shows a focus ring on keyboard focus; colours are CSS variable tokens; reduced-motion collapses the thinking dots to a static dot.

Palette tokens:#0b0c10#e6e8ee#22d3ee#a78bfa

Remix this brief: