DR
Projects/reelai
Live in ProductionAI / Content Automation2025 · Ongoing

AI Viral Content Automation

ReelAI

End-to-end multi-agent system that generates viral AI reels automatically: from news discovery to final video ready for TikTok and Instagram. Zero human intervention.

ClientInternal Project
RoleAI Engineer & System Architect
TimelineOngoing
Year2025
AI Viral Content Automation
The Challenge

Creating viral AI content for social media is time-consuming: research, scripting, recording, editing, captioning. Each reel takes hours manually.

Before

Manual process: find news, write script, record video, edit footage, add captions, export. Hours per reel with inconsistent quality.

  • Hours of manual work per reel
  • Inconsistent quality
  • Can't scale content production
  • Multiple disconnected tools
After

Fully automated pipeline: 7 specialized agents produce a publication-ready reel in 8-12 minutes for ~$4.70 USD. Zero human intervention.

Architecture

How it works

1

Discovery Layer

Perplexity API scans for trending AI news. Filters and ranks by virality potential.

2

Content Generation

Claude generates viral script. HeyGen creates avatar video. NanoBanana + Kling produce animated visuals.

3

Production Pipeline

FFmpeg extracts audio. Whisper transcribes. Remotion composes video with captions and transitions. Lambda renders in parallel.

Features

What was built

01

Multi-Agent Architecture

7 specialized agents with async execution and parallel processing. Each agent owns one step of the pipeline.

02

AI News Discovery

Perplexity API performs semantic search for trending AI news, feeding the content pipeline with fresh topics.

03

Viral Script Generation

Claude API generates optimized scripts with hook, tension, and CTA structure proven to drive engagement.

04

Digital Avatar Video

HeyGen API produces realistic talking-head video from the generated script. No camera needed.

05

AI Image + Animation

NanoBanana generates images, Kling AI animates them into cinematic sequences with motion and depth.

06

TikTok-Style Captions

Whisper.cpp transcribes audio locally, Remotion renders word-by-word animated captions with cinematic transitions.

07

Programmatic Video Composition

Remotion + React composes the final video programmatically. Lambda renders in parallel on AWS for speed.

Tech Stack

Technologies used

PythonPipeline Orchestration
Claude API (Sonnet)Viral Script Generation
Perplexity APIAI News Discovery
HeyGen APIAvatar Video Generation
NanoBanana ProAI Image Generation
Kling AICinematic Image Animation
Whisper.cppLocal Audio Transcription
Remotion + ReactProgrammatic Video Composition
Remotion LambdaDistributed Rendering (AWS)
FFmpegAudio Extraction
Results

Impact delivered

0Specialized AI agents
$0Cost per reel
~0minGeneration time
0Human intervention
Ready to build?

Let's build something together

Have a project in mind? Whether it's an AI chatbot, a web platform, or an automation system — I'd love to hear about it.