Skip to content
AspirecoStart
All answers

Automation & AI agents

How does automated short-form video production work?

The short answer

A pipeline runs the same stages every time — script, source footage, voice, render, publish, attribute — on a schedule, with a person reviewing drafts before anything reaches an audience. The hard part is not any single stage but keeping the whole thing running when one provider fails.

Updated

The six stages

Content Factory, our own pipeline, covers forty-two niches in four language regions:

  • Script — for time-sensitive niches, written against grounded search, so it cites the last 24 to 48 hours rather than model memory.
  • Source — B-roll through a licence-verified cascade of four ranked sources: stock, Creative Commons video, public archives and public-domain agency footage.
  • Voice — multilingual synthesis across the four language regions.
  • Render — composited with three fallback tiers per stage.
  • Publish — into a social scheduler as drafts.
  • Attribute — earnings tracked in a ledger that separates platform, affiliate and sponsorship rows.

Designed to degrade

One provider outage produces a slightly worse video rather than an empty schedule. Cosmetic steps fail open; irreversible steps fail closed.

A person before the audience

Everything lands as a draft, so a person stays between the pipeline and the audience.

Asked next

What tools does an automated video pipeline run on?

Content Factory runs on unremarkable parts — Node.js, Gemini, ElevenLabs, ffmpeg, n8n, Postiz, PostgreSQL and Railway. The value is in how the stages are chained and how each one fails, not in any single tool.

Can an automated pipeline publish in more than one language?

Yes. Content Factory runs across four language regions, with multilingual voice synthesis for each and scripts written per niche.

Read next

Start the audit