Minimax Hub 1.0 supports 3x3 storyboards and 5x15s clip assembly in creator tests

• Hailuo_AI's launch post framed MiniMax Hub 1.0 as a local AI agent creative workstation that runs research, script, image, music, and final-cut steps in one flow, while the official Hub site adds batch generation, local asset access, and export to pro tools.
• In DrSadek_'s character-sheet step, a Midjourney Niji creature became a six-view production sheet inside Hub, then DrSadek_'s follow-up split that design into a gear-free creature sheet and a separate detachable gear sheet for later shots.
• DrSadek_'s storyboard step shows Hub taking a short script and turning it into a shot list plus three 3x3 storyboards, which is the most concrete preproduction example in the launch-day evidence.
• The workflow still needed fallbacks: DrSadek_'s bug report says Hub video generation did not work in this test, so clips were made in Dreamina and the Hailuo app, then DrSadek_'s edit note says five 15-second clips were cut down and assembled in DaVinci Resolve.
• Audio was separate too. In DrSadek_'s music step, three storyboards were uploaded to Gemini to generate a 42-second background track, while Hailuo_AI's credit offer gave launch-day users 3,000 bonus credits before July 1 and said Hailuo credits transfer into Hub.

The Hub landing page promises script, storyboard, video, music, and editing on one canvas. In DrSadek_'s walkthrough, that promise turned into three 3x3 storyboards off a short horror script. A step later, the same thread reports Hub video was still buggy enough to require Dreamina and the Hailuo app. It still ends with a Gemini-generated 42-second score, which makes the launch feel half workstation, half relay race.

MiniMax launched Hub as a desktop creative agent product, not another single-model generator. Hailuo_AI's launch post pitches a chain from research to final cut, and the official Hub site spells that out as copy generation, image creation, video editing, audio and voiceover, auto packaging, and multi-format export.

The most concrete claims on the site are structural:

• Local-first workflow with access to local files and connected pro tools, per the Hub landing page
• Script, storyboard, video, music, and editing on one canvas, per the Hub landing page
• Four parallel agents for copy, image, video, and audio, per the Hub landing page
• Batch generation plus a shared local asset library, per the Hub landing page
• Manual checkpoints before the agent moves forward, per the Hub landing page

That pitch lands because it sounds like a preproduction desk, not a chat window.

The best launch-day proof came from a creator run that started outside Hub. According to DrSadek_'s character-sheet step, DrSadek_ first pulled a creature from Midjourney Niji, then uploaded it to Hub and asked for a character sheet.

The prompt visible in

is unusually specific about layout. It tells Hub to copy only the 3-column, 2-row grid from one reference image, preserve only the identity from a second image, and output one unified six-view production sheet instead of six separate pictures.

The next pass split the design into two reusable assets:

1. A clean creature turnaround with the removable tech stripped off, per DrSadek_'s gear-sheet follow-up
2. A separate gear layout with lenses, harnesses, canisters, pipes, and gauntlets isolated as standalone props, again in DrSadek_'s gear-sheet follow-up

That is a small but sharp workflow detail. The storyboard step later depends on having the body and the detachable equipment as separate references.

Once the sheets existed, Hub moved into preproduction. In DrSadek_'s storyboard step, DrSadek_ says a short script was uploaded and Hub was asked to improve it, suggest plate shots, and generate three 3x3 storyboards.

The visible prompt in

gives the scene in one paragraph: a post-apocalyptic underground lab, a beast escaping incubation, and half-dead zombies gearing it up. Hub's job was not only to draw frames, but to tighten the script and break it into shots.

This is the strongest creator-facing feature reveal in the evidence set because it shows a concrete intermediate artifact. Not just final video, but the planning layer between concept art and motion.

The handoff from storyboard to video was not seamless. DrSadek_'s bug report says MiniMax Hub was still new and buggy, and the video step failed in this run, so the storyboards were animated in Dreamina and the Hailuo app with a minimal prompt: animate this dark sci-fi film storyboard, no music, sound effect only.

The finishing workflow was still editor-led:

• Five 15-second clips were generated, per DrSadek_'s edit note
• Those clips were broken into individual shots and hand-selected, per DrSadek_'s edit note
• The final assembly happened in DaVinci Resolve, per DrSadek_'s edit note
• Film Look Creator added the final grain pass, again in DrSadek_'s edit note

So the current shape is clear enough. Hub can centralize planning and assets, but the motion layer can still spill into neighboring tools.

The last step in the walkthrough did not use MiniMax audio at all. In DrSadek_'s music step, DrSadek_ says three storyboards were uploaded to Gemini to create a 42-second background track, then the music was dropped into the edit at minus 12 volume.

That makes the launch-day pipeline look broader than the product pitch suggests. The creator kept Hub in the center for character sheets and storyboards, then borrowed Dreamina, the Hailuo app, DaVinci Resolve, and Gemini to finish the short.

A second early test from AllaAisling's first test called the new MiniMax Hub intuitive and fast, while stressing that the creator stays in control of the process. On the launch side, Hailuo_AI's download and credits post offered 3,000 bonus credits for logins before July 1, and said existing Hailuo credits transfer into Hub. That is a practical nudge for anyone already paying inside the MiniMax ecosystem.