ANDROS.BLANDON
← BACK TO INDEX/PROJECTS/SOCIALMAKER.AI
LIVE
YEAR
2025 — 2026
ROLE
FOUNDER, ENGINEER
STACK
NEXT.JS · CLAUDE API
STATUS
● IN DEVELOPMENT
CASE STUDY / 002 — PRODUCT

socialmaker.ai.

A SaaS platform exploring how generative AI can be operationalized responsibly for everyday creative work — turning AI from a curiosity into a tool people actually use.

[ TEMPLATE EDITOR — TO BE REPLACED ]
FIG. 01 — UNIFIED VISUAL TEMPLATE EDITOR
§ 01THE THESIS

Generative AI for content creation is a crowded category, but most tools fail in the same way: they treat AI as the product instead of as an ingredient. socialmaker.ai is built on the opposite premise — that AI is most useful when it's invisible inside a workflow that already makes sense to the person using it.

The platform pairs a unified visual template editor with prompt-driven content generation, allowing non-technical users to produce on-brand social media content at a pace that previously required either an agency or full-time creative staff.

§ 02ARCHITECTURE

socialmaker.ai is a multi-tenant SaaS built on Next.js 15 (App Router), React 19, TypeScript, and Prisma ORM with PostgreSQL persistence. The visual template engine renders editorial layouts as JSON-described scenes, supporting freeform text layers, decorative elements, and programmatic PNG export.

Generation requests route through the Anthropic Claude API, with prompt engineering tuned per template archetype. The system separates content-generation responsibilities from rendering responsibilities, allowing model upgrades and template iteration to proceed independently.

§ 03ENGINEERING NOTES

The visual rendering system started as separate freeform and structured renderers, then unified into a single editorial renderer with template-described layer logic. This consolidation simplified the export pipeline and made variant selection deterministic.

MCP (Model Context Protocol) integration is being explored as the mechanism for connecting socialmaker.ai's templates to user-owned brand asset stores — allowing the AI to reason about a user's brand context the same way a human designer would, without leaking private information into shared model contexts.

§ 04STATUS

The MVP is functional and in private development. Public launch is planned for late 2026, with initial focus on a small group of design partners providing real workflow feedback before broader release.