HKUDS

ViMax

HKUDS multi-agent video studio: turns an idea, novel or screenplay into a finished film — scriptwriting, storyboards, consistent characters, then rendering via Seedance/Nano Banana/Omni APIs.

12,121 1,827 Python MITupdated 1 months ago

vision agents

View on GitHubDispute this mapping →

Curator's take

Use it when you want long-form narrative video rather than 8-second clips: the agent pipeline (screenwriter → storyboard artist → character extractor → portrait generator → shot renderer) is engineered for cross-shot character and scene consistency, with an interactive agent-loop TUI for revision, resume and context compaction. Novel2Video — compressing a whole book into episodic video — is the flashiest trick. NOT a video model: you bring API keys for the generators (Doubao Seedream/Seedance, Nano Banana, MiniMax, Google Omni; several routed via the yunwu.ai reseller, so costs are real and defaults skew CN). Same lab as VideoAgent — pick VideoAgent to work on existing footage, ViMax to create from scratch.

Mapped by ShipWithAI editors · links verified

README.md

(no README fetched)

Continue your stack

What teams reach for next — and why each earns a place beside ViMax. Ranked by curator confidence.

Pairs well with

video-spec-builder

A director-skill for Claude Code/Codex: interrogates your vague video idea until it's a second-by-second storyboard spec (video-spec.md), ready for HyperFrames to render.

Why it fitsSpec-first meets render-pipeline: video-spec-builder interrogates a vague idea into a second-by-second storyboard spec, exactly the kind of screenplay input ViMax's Script2Video pipeline consumes (no official integration — the spec skill targets HyperFrames).

Alternative to See all + compare →

VideoAgent

All-in-one agentic framework for video: understanding and summarization, clip editing, and generative remaking, driven end-to-end through natural-language conversation.

Why it fitsSame lab (HKUDS), overlapping agentic-video job, opposite direction: VideoAgent operates on existing footage — understand, summarize, edit, remake — while ViMax generates films from an idea, novel or screenplay from scratch.

hyperframes

Write HTML, get MP4: HeyGen's agent-native video framework renders deterministic GSAP/Puppeteer/FFmpeg compositions, shipping 20 skills and MCP so coding agents author motion graphics.

Why it fitsSame ask — 'make me a video' — opposite mechanics: Vimax scripts and generates footage through Seedance/Nano Banana models, HyperFrames renders HTML you control frame-exactly. Vimax for cinematic content you can't hand-author, HyperFrames when precision and repeatability matter.

MoneyPrinterTurbo

Topic in, finished short out: an LLM writes the script, footage comes from Pexels/Pixabay or text-to-video, then TTS, subtitles and music compose into HD 9:16 or 16:9. WebUI, API and CLI.

Why it fitsSame pipeline shape — script, visuals, assembly — at opposite ambitions. Vimax targets films from novels or screenplays with character consistency; MoneyPrinterTurbo targets the 60-second social short you can batch a dozen of before lunch.