I built an open-source, offline-first AI Content Factory — one topic in, a complete content package out
Most AI content tools solve one step.
Write an article.
Generate an image.
Make a video.
Create a social post.
I wanted to build the pipeline around the entire workflow instead.
So I built AI Content Factory, an open-source, provider-neutral content production pipeline.
Give it a topic, and the system can turn it into:
Topic
↓
Research
↓
Article
↓
Short-form script
↓
Storyboard
↓
Image / video / voice plans
↓
Media QA
↓
Approval & integrity checks
↓
Platform-ready content packages
↓
A visual HTML preview
The important part is that the core is not tied to a specific AI company, model, or paid API.
The default v0.1.0 demo runs completely offline.
No API key.
No account.
No dedicated GPU.
No paid service.
No private assets required.
You can clone it, run the demo locally, inspect every artifact it creates, validate the output, and then replace the included deterministic providers with your own image, video, voice, research, or LLM adapters.
The core orchestration only knows contracts — not vendor SDKs.
That means the provider layer can evolve without rebuilding the whole content workflow around a single model or service.
I also deliberately kept the first public release conservative:
The goal of v0.1.0 is not to pretend that an offline demo magically produces production-quality AI media.
The goal is to make the orchestration, packaging, safety boundaries, and provider architecture reproducible first — and then let people plug in the models and services they actually want to use.
The project is written in Python and released under Apache-2.0.
GitHub:
[https://github.com/kafidog/ai-content-factory](https://github.com/kafidog/ai-content-factory)
If you’re working on local AI, content automation, multimodal pipelines, provider adapters, or reproducible agent workflows, I’d especially like to hear what feels useful — and what you’d replace first.
If you’re working on local AI, content automation, multimodal pipelines, provider adapters, or reproducible agent workflows, I’d love to hear what you’d replace or improve first.
If you find the project useful, consider giving it a on GitHub — it helps more people discover the project.