AWS + AXUM High-Margin Visual Syndication Engine

Transform Viral Trend Signals into Automated Video Syndication

ShareCaster Studio replaces raw media ingestion with dynamic social trend harvesting. Powered by a high-throughput Rust engine, 2-tier signal deduplication, and self-correcting generator-validator agent loops.

Test Live Event Stream Inspect Tech Spec

95.4%

Low-Signal Noise Dropped

< 12ms

Tokio SSE Dispatch Latency

Zero-Bloat

Rust / SQLx Architectural Core

Autonomous

Background Co-Pilot Cron

Ingestion & Curation Engine

Two-Tier Trend Curation Pipeline

Social feeds are full of noise. We discard 95%+ of low-signal traffic using fast Rust heuristics before routing high-velocity topics to lightweight classification nodes.

Status: Filtering Active Feeds

Tier 1 Node Rust Heuristics

Deduplication & Velocity

Fast regex matching, duplicate hashing via SQLx PostgreSQL state, and engagement velocity calculation.

Velocity Threshold 750 score/hr

Dropped Noise: 82.4%

Tier 2 Node Lightweight 8B Model

Context & Narrative Check

Low-latency classification passes to ensure story depth, trend authenticity, and visual adaptability.

Model Inference Route DeepSeek-8B-Fast

Passed (88%) High Cost LLM Skipped

Total Curation Yield: 4.6% High-Signal Payload

Stage Output Ready for Agent

Clean Signal AST

// Parsed JSON AST Payload

"trend_id": "tr_98421_ai"

"virality_score": 94.8

"hook_vector": "tech_breakthrough"

Self-Prompting Meta-Engine

Generator-Validator Agent Loops

Static prompt engineering is fragile. ShareCaster uses an asynchronous dual-agent loop: a Generator node drafts the story layout, while a secondary Validator inspects AST dynamic constraints, automatically prompting delta retries until verified.

Signal Payload

Raw vetted trend AST received from Tokio background worker queues.

Script Generator

Synthesizes visual storyboard frames, voiceover timing, and video hook variations.

AST Validator

Dynamic verification against strict JSON schemas, pacing rules, and layout logic.

Feedback Delta Loop

S3 Render Dispatch

Verified payload sent to AWS Lightsail/S3 rendering workers for final media assembly.

Live Axum Server-Sent Events

Interactive Studio Engine Simulator

Experience real-time Tokio thread dispatching, multi-agent validation loops, and background cron pipeline updates streamed live to the client UI.

Creator Decision Tree

Zero Typing Onboarding

Niche / Signal Feed:

Target Video Format:

Co-Pilot "Ready-to-Render"

Cron Auto-Drafted

Quantum Computing Breakthrough Hook

Score: 98.2 • 3 Stories Ready

Rust Axum 3.0 Architecture Overview

Score: 94.1 • Ready for Preview

sse_stream_listener.rs — api.sharecaster.com/events

SSE CONNECTED

// ShareCaster Axum SSE Receiver Initialized...

// Waiting for background cron payload or manual trigger...

Active Tokio Threads: 8

S3 Bucket Target: s3://sharecaster-media-prod/

Systems Architecture

Engineered for Performance & Scalability

Zero-bloat dependencies, native Rust speed, and serverless edge distribution on AWS.

Axum + Tokio Backend

Asynchronous Tokio task management ensuring near-zero HTTP latency, safe multi-threading, and rock-solid database access via SQLx PostgreSQL connection pooling.

AWS Managed Cloud

Route 53 DNS routing, restricted IAM role boundaries, automated S3 storage buckets, and high-speed Lightsail CDN distribution edge delivery.

React & Vite Dashboard

Clean single-page creator workspace driven by real-time SSE event hooks, multiple-choice onboarding flows, and instant video previews.