ClipNest
My own AI product: it writes, narrates and renders short videos. The same pipeline work goes into AI features for clients.
- AI product, my own
- Invite-only beta
- 2025
- Everything: product, build, hosting
What it does
A video channel run by a pipeline
You describe a channel. ClipNest researches the topic, writes the script, records the narration, picks matching footage and renders the video, then does it again on a schedule.
How a video gets made
Five stages, with checkpoints between them
Research
An AI model gathers what the video should say.
Script
Outline, script and hook, checked between steps.
Narration
The script becomes a voice track.
Footage
Each line of the script is matched to fitting clips.
Render
The video is cut and rendered, then scheduled.
Why it matters for your project
AI features here are production work, not experiments
Everything I offer as AI automation on client projects, from content pipelines to AI-backed features and background processing, I have already shipped and run in ClipNest.
Picks up where it stopped
The script stage resumes from its last checkpoint instead of starting over.
Cost control
A monthly cloud budget with an automatic kill-switch keeps the bill predictable.
Retries where they're safe
A failed render reruns on its own; the paid AI steps run once, on purpose.
Built with
NestJS, BullMQ job queues, AWS Lambda for rendering, Terraform for storage, the footage service and the spending cap.
