GenJam 2026 — Engineering Track

The algorithm hearseverything.

Add a punchlyne to your video—one sentence at the end that platforms transcribe, index, and use to find your audience.

Try the demo

The Problem

Platforms decide who sees your video based on what they hear.

YouTube, TikTok, Instagram

Every platform transcribes your video. That text decides what it's about and who should see it.

The creator's dilemma

Shoehorning a sponsor mention mid-video feels awkward. The content suffers, and viewers leave.

The Concept

One sentence at the end. Your mouth. Their algorithm.

PunchLyne adds a second clip—a punchlyne—where your own face delivers one casual sentence containing the search phrase. The original video stays untouched.

A viewer might leave after the opening, but the platform transcribes every second. The punchlyne gets heard.

How It Works

Five steps to a punchlyne.

  1. 01

    Upload or record

    Upload an existing talking-head video or record a few seconds of your face.

  2. 02

    Describe the target

    Say who the video is for and what you're lowkey plugging. PunchLyne does not guess.

  3. 03

    Generate the line

    AI writes one casual sentence that includes the search phrase, then speaks it.

  4. 04

    Lip-sync

    The lip-sync API moves your mouth to that audio and returns a separate short clip.

  5. 05

    Take it

    Download both clips, or staple them into one file ready for upload.

Tech Stack

Built for the Texel track.

LayerTool
FrontendReact 19, TanStack Start, Tailwind CSS
HostingVercel
Line generationxAI (Grok)
Voice synthesisxAI TTS
Lip-syncTexel API
Video joinTexel Video Encoder

Demo

Try it now.

1. Your video

Upload it, or record a few seconds of your face.

Or try a sample face, so you can see the clip without filming.