Why Pip?
A curious-facts Shorts channel run by an automated AI pipeline, from trending topic to daily upload.
File summary
- Filed under
- Media AI, Automation
- Built by
- Varun Chennuri, with an AI engineering team
01
The problem
A daily Shorts channel is a daily production line: a sourced topic, a script, art in one consistent style, a voice, an edit synced to the words, a thumbnail and an upload. Every single day.
Why Pip? hands the whole line to an automated AI pipeline.
02
What I built
One automated pipeline, from trending topic to upload.
- Picks a daily trending topic, with strict sourcing rules.
- Writes the script.
- Generates sketch-style images in one consistent style.
- Voices it with ElevenLabs v3, using emotion tags.
- Syncs every picture to the exact spoken word with a forced aligner.
- Renders with ffmpeg.
- Checks pronunciation with Whisper.
- Makes the thumbnail.
- Uploads to YouTube and Instagram on a daily schedule.
03
How AI agents built it
Here the AI agents are the production team.
- Checks are built into the line: strict sourcing rules before a topic is picked, word-level timing from a forced aligner, and a Whisper pronunciation check.
- One visual style for every image, so the channel looks like one hand drew it.
- The same pipeline makes brand explainers for Karkaana and Cyepro.
04
Architecture sketch
Drawn the way it sits on my desk: boxes for the moving parts, arrows for the flow, dashed lines for what is wired in alongside.
- 01TRENDING TOPICpicked daily, strict sourcing rules
- 02SCRIPT
- 03SKETCH IMAGESone consistent style
- 04VOICEElevenLabs v3 + emotion tags
- 05FORCED ALIGNEReach picture on its spoken word
- 06RENDERffmpeg
- 07PRONUNCIATION CHECKWhisper
- 08THUMBNAIL
- 09YOUTUBE@whypip
- 10INSTAGRAM@whypip.daily
05
Results
Real numbers only. Anything not confirmed yet is marked as a placeholder.
- 69Shorts produced since late October 2026Basis: Channel output
- Dailyupload schedule on YouTube and InstagramBasis: Pipeline schedule
- 2 brandsKarkaana and Cyepro also get explainers from itBasis: Pipeline output
06
Stack
- ElevenLabs v3
- Forced aligner
- ffmpeg
- Whisper
- YouTube
Next step
Want this for your company?
Start a 7-day sprint and get a working AI product: two AI models that review each other, tests on every change, and me on every risky call.