Scaffold AI Platform
A production-grade, enterprise multi-agent orchestration and visual creative AI playground. Built with an extensible node-graph DAG engine (@xyflow/react + LangGraph), multi-modal cloud AI pipelines (FAL AI Flux/Luma/Wan + Google Gemini LLMs), real-time credit consumption ledger, and hybrid Supabase JWT + FastAPI RBAC security.
View GitHub repoVisual Node-Graph DAG Engine & Architecture
Scaffold bridges high-throughput asynchronous AI execution with an intuitive visual graph interface. Users construct, chain, execute, and monitor multi-modal workflows combining LLM text generation, image synthesis, video generation, and audio processing.
Backend System Topology
The backend microservices layer is powered by FastAPI, AsyncIO, and Celery task workers backed by Redis and Supabase PostgreSQL.
Each node execution calculates credit costs dynamically, verifies the user's available credit balance via atomic transactional logs, invokes cloud providers (Google Gemini or FAL AI), and streams real-time status updates back to the visual canvas.
Core Platform Capabilities
Desktop & Web Application Gallery
Explore all key desktop interfaces, blueprint visual editors, and admin management portals of the Scaffold AI Platform.
Mobile Responsive View Showcase
Optimized mobile touch interfaces for monitoring agent workflows, blueprint graphs, and credit consumption.
Technology Stack & Infrastructure
| Backend Framework | Python 3.12, FastAPI 0.109+, AsyncIO, SlowAPI Rate Limiter |
|---|---|
| Frontend Framework | Next.js 16, React 19, Tailwind CSS v4, `@xyflow/react`, Dagre Auto-Layout |
| Database & State | Supabase PostgreSQL (UUID v7, JSONB, RLS), Redis 7 Broker & Celery Result Backend |
| Cloud AI Models | Google Gemini 2.5/3.0, FAL AI (Flux.1, Luma Dream Machine, Wan 2.1, ElevenLabs), LangChain/LangGraph |
| Vector & Asset Storage | Pinecone Vector DB (Semantic Asset Embeddings), Supabase Storage CDN |









