01
What is DEVLAR?
An autonomous media pipeline that discovers developer news, applies two-pass Gemini AI editorial judgment, and renders + uploads short-form videos — hands-free.
4
Data Sources
HN · GitHub · Product Hunt · RSS
100%
CPU Saturation
All cores pinned on render
3
Platforms
YouTube · Instagram · Facebook
0 B
Cloud Storage
Ephemeral S3 bridge
Interactive Architecture
The 5-Stage Pipeline
Click any stage to inspect its inputs, mechanisms, and outputs in detail.
STAGE 014 Independent Streams
Content Discovery
Runs four independent collectors concurrently — failure in one never blocks the batch.
4 collectors
Hacker News APIJSON REST
Polls top stories with score & discussion velocity.
Engineering BlogsRSS / Atom
Direct feeds from OpenAI, Google DeepMind, Meta, AWS.
GitHub ReleasesREST API
Major version releases from tracked infra repos.
Product HuntRSS Feed
Developer tools, AI infra, and productivity launches.
Navigate the Docs
Explore the System
Four core questions — click through to the full documentation page for each.
Read the philosophy
02
How does the pipeline work?
Five deterministic stages: collect → filter → AI newsroom → render at 100% CPU → simultaneous broadcast to YouTube, Instagram, and Facebook.
03
What technologies power it?
Python collectors, Gemini Flash editorial brain, Edge TTS neural voice, Remotion multi-core renderer, AWS S3 ephemeral bridges, and Next.js + MongoDB Atlas for presentation.
04
How did it evolve?
Started as a GitHub scraper prototype. Grew through 7 architectural milestones into a triple-platform cloud media system deployed on AWS EC2.
See It Running Live
The pipeline runs autonomously on AWS EC2. The Control Center shows real-time publishing stats and performance telemetry from each batch.