Industry
Content Tools
Case Study
Story Content Generator
Producing a story sequence should not mean rebuilding every frame by hand. I built a guided desktop workflow that turns structured copy and brand settings into reviewable 9:16 frames, with optional local AI drafting and a deterministic fallback.
Project Overview
Social stories are simple to consume but repetitive to produce. Every frame needs consistent typography, spacing, brand rules, safe zones, and exact 1080 by 1920 output.
I built a local desktop tool that turns an initial brief into an editable sequence, live preview, and final PNG export without moving text or logos into an image-generation model.
A local desktop editor for drafting, reviewing, and exporting deterministic 1080 by 1920 story sequences from structured content.
Services
- UX/UI Design
- Web Development
- Product Strategy
Key Deliverables
- Responsive Design
- Custom Development
- User Experience Improvements
Platform Type
Desktop Tool
Tech Stack
- Python
- FastAPI
- Playwright
- PyWebView
- Pillow
The Challenge
Manual design makes even a small copy change expensive. Fully generated images are also unreliable for precise text, logos, hierarchy, and Instagram interface safe zones.
A useful tool must accelerate ideation while still producing the same layout in preview and export, even when no AI service is available.
Approach
A guided wizard collects the input and writes readable content.md plus project overrides. Story archetypes such as hook, tip, insight, statistic, list, quote, myth/fact, and CTA give each frame an explicit role.
FastAPI serves the local UI inside PyWebView. Jinja2, HTML, and CSS produce one 9:16 renderer that Playwright uses for both preview and export.
The Solution
Users can create a sequence, edit frame copy, move between frames, tune visual options, preview the result, and export PNGs from one desktop workflow. FontTools and Pillow support typography validation and final image processing.
If a local Ollama instance is available and AI drafting is enabled, it can propose the first story sequence. When it is unavailable, the deterministic planner builds a valid sequence instead, so the core tool never depends on the model.
Key Challenges
-
keeping preview and exported images visually identical
-
fitting variable-length copy inside strict vertical safe zones
-
creating a useful sequence when the optional local model is unavailable
-
separating assisted drafting from deterministic brand rendering
The Outcome
The result turns repeated story production into a short draft, edit, review, and export loop. AI can accelerate the first draft, but the content structure, brand elements, and final pixels remain inspectable and deterministic.
Context
The project uses AI as an optional starting point, not as a runtime dependency or final renderer.
Related case studies
Other projects with related architecture decisions, platforms, or product challenges.
Social Carousel Generator
A cross-platform desktop pipeline that turns structured content and brand rules into deterministic carousels, posts, and highlight graphics.
Python · FastAPI · Playwright · PyWebView · Pillow
View case study →Web Screenshot Tool
A Bun and Hono tool that discovers pages through sitemap traversal, captures 1920-pixel full-page screenshots, and streams live progress to a lightweight TypeScript interface.
TypeScript · Bun · Hono · Puppeteer · Server-Sent Events
View case study →