A multi-tenant command center for social video: an idea becomes a script, the script becomes a cinematic AI-generated reel, the reel gets per-platform captions, and the whole thing schedules out to Reels, TikTok and Shorts. Currently runs in mock and dry-run mode — nothing is published and no credits are spent at rest.
The problem
Running short-form video for several brands means repeating the same chain — idea, script, footage, captions, per-platform copy, scheduling — over and over, with the context for each brand held in someone's head.
What I built
Verified end to end in mock and dry-run mode: the full chain runs, produces its artefacts, and stops short of publishing. Deliberately left unarmed until the pipeline is trusted.
How it works
The core of the system is a prompt engine that turns a brief into structured, preset-driven generation instructions rather than freeform text, so output stays consistent across runs and across brands. A Python API owns generation and scheduling; a separate React front end is the operator's console. Every external side effect is behind a live/dry-run flag that defaults to off.
Highlights
- Multi-tenant — one operator, several brands, isolated context each
- Preset-driven prompt engine for consistent generation instead of freeform prompting
- Every publishing side effect behind a dry-run flag that defaults to off
- Verified end to end without spending generation credits or posting anything
Let's build yours.
Tell me what you're trying to ship — you'll get a scoped plan and a straight answer.