AI Filmmaker · Est. on X

PIXELAIGC

二 十 分 钟 A I 长 片

I make 20-minute AI films — complete, unabridged adaptations of classic Chinese literature, built frame by frame with Midjourney, Seedance and MiniMax. Not clips. Whole stories.

10,000+
Followers on X
50,000+
Views every day
32
Films published
133 min
Of finished film
01

Featured Films

代 表 作

Every film is written, directed, generated and finished in-studio — full production notes are shared with subscribers on X.

02

Track Record

资 历

Midjourney pioneer

Among the earliest image-generation creators anywhere — years of prompt craft sit behind every frame of every film.

Creative Partner CPP

Official partner creator of four major AI video platforms: MiniMax, Vidu, Luma and PixVerse.

Seedance 2.0 classics

A body of classic works from the platform's first wave — including a hotel fight film so convincing that viewers refused to believe it was AI. He later broke it down prompt by prompt in a public tutorial article.

READ THE BREAKDOWN →
THE ORIGINAL FILM →

MiniMax H3, in depth

Deep, production-level research on H3 — from local ComfyUI deployment and Turbo LoRA tuning to true long-form filmmaking. The open breakdown post drew 80K views and 675 bookmarks; the workflow is open-sourced on GitHub — ComfyUI-MiniMax-H3-Workflow (86 stars, 19 forks).

READ THE POST →
THE 442K-VIEW H3 HIT →

Dozens of awards

Works awarded again and again across AI video competitions — craft recognised by juries, not just feeds. Anchor credential: Second Prize at the 2024 Wenzhou AIGC Video Creation Competition (251 entries, official winners list).

OFFICIAL WINNERS LIST →

Full-stack fluency

Midjourney · Seedance · MiniMax · Vidu · Luma · PixVerse · Muse — one workflow, every major model, chosen per shot.

Creative Partner ofMiniMax·Vidu·Luma·PixVerse
Also working withMidjourney·Seedance·Muse·Pika
03

Studio Stack

技 术 栈
The production pipeline
Midjourneykey imagery
→
ComfyUI · MiniMax H3video generation
→
Turbo LoRA6-step sampling
→
Topazupscale
→
Segment assemblyup to 130 parts
→
One-pass narrationuniform speed
GPU ...... RTX 4070 · 16GB VRAM · local
ENGINE ... ComfyUI · H3 · Seedance
SAMPLE ... Turbo LoRA · 6 steps
CFG/SHIFT 1.0 / 6 · 768p native
OUTPUT ... films up to 22 min · 16:9
SOURCE ... github.com/pixelaigc ★86
04

For Brands

品 牌 合 作

What I can do for you

  • Sponsored posts & quote posts for AI models and tools
  • Original AI videos showcasing your product
  • Model-launch campaigns with technical depth — I read the paper, not just the press release
  • Long-form AI film production & consulting
Recent collaboration

Pika — H3 Flex Launch

Paid partnership for the H3 Flex model launch (Nov 2026): an in-depth quote post built from Pika's own technical blog — training pipeline, distillation, speed benchmarks — with the blog's charts, plus community engagement across my channels.

SEE THE CAMPAIGN POST →
05

About

关 于

PIXELAIGC is a one-person AI film studio. While most AI video lives in 10-second clips, this studio works at the opposite end of the craft: complete literary works adapted into finished films of up to 22 minutes — character-locked, narrated in one pass, and quality-checked frame by frame.

The audience is filmmakers, AI builders and literature lovers across the Chinese-speaking world and beyond — the exact crowd that adopts new models first.

Brand collaborations & commissions