Personal television, generated in real time

Your taste, broadcasting.

footage.center turns the people, ideas, music, and quests already shaping you into an endless channel of directed short-form video. Every clip is yours. Every clip is a real production.

Graph ranked3 to 10 clips aheadStudio one tap away

Station tuned from 428 taste signals

Interactive prototype
Personal signal97% match
FC · 1842

The city that remembers every song

city pop × memory architecture

Every street kept one chorus from the person you were.

00:0400:24
01

The graph programs.

footage.center reads follows, saves, watch time, people, topics, music, and the quest you are on now, then ranks which idea deserves the next slot in your personal channel.

02

The director composes.

The footage.center director applies a fixed episode grammar, turning each taste match into a deliberate 12 to 30 second story with voice, captions, and music instead of a random clip.

03

The queue stays ahead.

The footage.center queue is designed to render the next 3 to 10 clips while you watch. Model adapters can change without touching your channel, production record, or viewing history.

04

Your actions retune it.

Skip, rewind, remix, and save update the footage.center taste graph immediately, so the next prefetched episode changes with you rather than with a generic trend chart.

Episode grammar

Infinite does not mean shapeless.

Every footage.center generation must earn four beats: opening texture, human detail, system in motion, and closing image. The structure is stable; the worlds, characters, and visual language can keep changing.

00:00

Opening texture

A sensory signal that places you inside the world before it explains itself.

00:04

Human detail

One person, gesture, or line that makes the episode emotionally legible.

00:11

System in motion

The idea becomes visible through action instead of exposition.

00:23

Closing image

A final frame with enough charge to live in memory and seed the next clip.

The durable layer

The station owns direction. Models only render.

footage.center keeps the interest graph, episode brief, source ledger, edit decisions, consent, rights, feedback, and final renders. The fastest suitable model gets the shot, then gets out of the way.

Read the product contract
PeopleIdeasMusicQuest
footage.centerGraph + director + queueProduction record remains canonical
Reels 9:16StoriesAdsFilms

Status and access

A free prototype, honestly labeled.

footage.center is an early, free browser prototype built by Adam Pang for people with a rich digital taste trail who want a feed shaped by their own circle, ideas, music, and current quest. There is no pricing, plan, account, or checkout yet.

The working parts of footage.center today are Cable playback, feedback controls, taste weights, queue state, and Studio records in the browser. Interest graph ingestion, prefetch generation, and the video model adapter are still simulated, so faster-than-realtime generation is not yet a proven claim.

Not another prompt box

Press play. Meet the channel only you could have.

Try the prototype