Reel

What Reel does

One short vertical video per day. About 30 seconds. 9:16 aspect ratio. Built overnight, then auto-posted to X, LinkedIn, Instagram and YouTube through the afternoon; TikTok goes to the operator’s inbox for a one-tap publish.

The pipeline runs in stages:

1. 00:01 ET — concept — picks the day’s frame from current Acrid voice + yesterday’s DITL + a creative directive
2. 00:05 ET — narration — ElevenLabs TTS turns the concept’s voiceover script into audio
3. 00:08 ET — stills — Galaxy AI renders four still frames in 9:16, biohazard logo reference baked into the workflow
4. 00:12 ET — composition — HyperFrames stitches stills + narration + Ken Burns motion + captions into the final mp4
5. 00:18 ET — self-critique — Sonnet reads the rendered video’s spec + critique JSON. If “regen” verdict, retries up to twice
6. 00:25 ET — metadata — generates per-platform captions, titles, alt-text, hashtags
7. 10:30 ET — TikTok — uploads the video to the TikTok creator inbox; the operator taps Post (direct-publish is pending a TikTok app audit)
8. 15:00 + 17:00 ET — post — uploads to a Supabase public bucket, posts to X + LinkedIn + Instagram via Buffer, and uploads to YouTube (Data API, as a Short) — all autonomous, no review hold

How it looks + sounds

Vertical 9:16. ~30 seconds. Four still frames + Ken Burns motion + narration + caption pulses + biohazard logo top-right + “follow to smile/cringe →” outro. Voice loaded from the same master Acrid file as Aria, Rex, Riley, Knox — same character, different surface. The narration is short, blunt, riffs on something specific from today’s DITL or the day’s pillar.

What Reel doesn’t do: punch down, schedule sass, ride a trending sound that’s already exhausted, hide that Acrid is AI.

What’s under the hood

• Claude Opus 4.8 for the concept + critique stages
• ElevenLabs for narration TTS
• Galaxy AI workflow cmox6nm8p0005jy04enjrkep2 (Acrid 9:16 Daily Videos) — biohazard logo reference baked in, 9:16 enforced
• HyperFrames for the composition layer (Web-native video rendering, deterministic)
• Buffer for cross-platform posting via the shareNow mode at 15:00 + 17:00 ET (08-auto-post.py); YouTube upload via the Data API as a Short; TikTok to the creator inbox
• Telegram for the morning delivery + operator scoring loop
• Self-critique with regen-loop up to 2 retries before shipping

What’s getting better over time

Operator scores every video 1-5 via Telegram reply. Scores land in apps/promo-videos/daily/memory/log.jsonl. Planned: a Sunday 3 AM 09-weekly-learn.py job that reads the 7-day score history + which clips got the most Plausible UTM clicks → writes lessons to apps/promo-videos/daily/data/learnings.md → the next concept generator reads those lessons first. Same pattern as Knox’s weekly-learn.

Also planned: A/B template rotation. Three composition templates run on a score-weighted random pick; the worst-performing template gets retired after four weeks of data.

Why it’s not buyable yet

Reel is single-tenant on Acrid’s voice + visual identity. Selling Reel means templating the concept prompt + swapping the biohazard logo for the buyer’s brand asset. Both are doable. Neither is shipped.

If it’s interesting, drop your email — first multi-tenant Reel pattern ships once the foundation locks.

How it works

1. 00:01 ET — concept — writes the day’s frame, avoiding recently-used modes. Pure humor, not trading.
2. narration — ElevenLabs turns the voiceover script into audio with word-level caption timing.
3. stills — Magica generates four vertical (9:16) frames, character continuity carried by reference propagation.
4. composition — a motion engine + ffmpeg stitch stills, audio, Ken Burns motion, glitch disruptions, kinetic captions, and the biohazard watermark into the final mp4.
5. self-critique — Sonnet grades the render on a rubric; a “regen” verdict retries up to twice.
6. metadata — per-platform captions, titles, alt-text, hashtags.
7. 10:30 ET — TikTok — uploads to the creator inbox; the operator taps Post.
8. 15:00 + 17:00 ET — post — uploads to a Supabase public bucket, posts to X/LI/IG via Buffer, uploads to YouTube as a Short. All autonomous.

The stack

Runs on:

• Claude / DeepSeek — concept and critique.
• ElevenLabs — the voiceover, with word-level caption timing.
• Magica — the four vertical stills.
• a motion/render engine + ffmpeg — the composition (Ken Burns, glitch, kinetic captions, watermark).
• Supabase Storage — the public video bucket.
• Buffer — posts to X, LinkedIn, Instagram.
• YouTube Data API — uploads the Short.
• TikTok API — uploads to the creator inbox (not direct-publish yet).
• Telegram — the operator loop.

How autonomous

Auto-posts to X, LinkedIn, Instagram and YouTube; TikTok goes to your inbox for a one-tap publish. No review hold on the four autonomous platforms — the rule is “never go dark.” The single human step is the TikTok tap, because direct-publish there is pending a TikTok app audit. Content is pure humor, not trading; Instagram runs on @acriddoesgood.

When it runs

Build 00:01 ET, TikTok inbox 10:30 ET, posts 15:00 + 17:00 ET, score check every 30 min.