Inam Baig / projects / content-os
AI / Content Automation
Content OS
Multi tenant OS that discovers topics, generates scripts with LLMs, and produces short form video.
100% developed with agentic engineering.
Role
Agentic Engineer · Product Manager · Technical Lead
Duration
… to Present
Domain
Automated Intelligent Video Generation · Worldwide
Tech highlights
Problem
What was hard
Overview
Running many short form channels requires repetitive discovery, writing, editing, and publishing work.
Users affected
Creators, studios, agencies operating multiple niches.
Why difficult
End to end pipeline across AI, assets, render, and publish without forking per niche.
Existing limitations
- Point tools don't cover the full lifecycle
- Hardcoded niches don't scale
Constraints
- Multi tenant
- Human approval before publish
- Deterministic first AI
Solution
Approach
Overview
Content Sets + plugins with FastAPI, Next.js, Postgres, Redis, RabbitMQ, Remotion, Playwright, OpenAI.
Architecture rationale
Niches as configuration; LLMs only for narrow schema validated tasks.
Major components
- Web app
- API
- Workers
- Renderer
- Queues
- Storage
Key capabilities
- Discovery
- LLM scripts/scenes
- Remotion render
- Publishing
- Analytics feedback
Contribution
My contribution
Project outcome
I designed and built a multi tenant intelligent video generation platform.
What I personally owned
- I owned 100% of product and engineering with agentic engineering
- I owned SRS, architecture, API/web/pipeline implementation
Features
Key capabilities
Expand a card for implementation detail and user value.
Architecture
Interactive diagram
Zoom, pan, and select nodes to explore components and connections.
Operators drive Content Sets from a Next.js app into a FastAPI orchestrator. RabbitMQ and Redis coordinate Python workers that call OpenAI, crawl with Playwright, render with Remotion/FFmpeg, and persist state in PostgreSQL plus S3/MinIO object storage.
Stack
Technology
Grouped by layer. No skill bar percentages.
Frontend
- Next.js
- React
- TypeScript
Backend
- Python / FastAPI
- Background workers
Database
- PostgreSQL
- Redis
Infrastructure
- RabbitMQ
- S3 / MinIO
- Docker
AI / ML
- OpenAI API
Other
- Remotion
- FFmpeg
- Playwright
Challenges
Engineering challenges
Automating many short form niches without forking the codebase
Why hard
Point tools don't cover discovery → script → render → publish, and hardcoded niches don't scale.
Solution
I modeled niches as Content Set configuration. FastAPI orchestrates workers; LLMs only fill narrow schema validated script/scene tasks; humans approve before publish.
Result
One deployment can run many channels with Playwright discovery, Remotion/FFmpeg render, and RabbitMQ/Redis pipelines.
LLM cost and quality control
Why hard
Unconstrained generation is expensive and drifts off brand.
Solution
Deterministic first pipeline: OpenAI only for schema validated generation, Content Set tone/niche as config, human approval gate before publish.
Result
Generation stays bounded and reviewable instead of a free form agent loop.
Results
Outcomes
Outcomes
- Documented SRS and local runnable stack
Related
Related projects
Also look at
