VibePaper: A Short-Drama AI Workbench That Puts Four Agents on One Canvas

The short version

On most canvases you work alone; on VibePaper's canvas you work alongside an AI team. It puts planning, screenwriting, visual and editing agents on the same node-based infinite canvas, driven by leading language, image and video models, closing the loop from script to finished cut. The AI executes; taste and key decisions stay with your core team.

What VibePaper is

VibePaper is an AI collaboration workbench aimed at short-drama production companies and professional creators. It supports AI comic dramas, human-like short dramas, commercial spots and more, all within a single canvas — no shuttling footage between prompt tools, image tools, video tools and editors.

Unlike conventional single-point tools that do nothing more than "type a prompt, get a video," VibePaper offers a complete workflow from brief to finished film. It breaks creation into independently editable nodes — script, storyboard, shot, voiceover — so changing one step reruns only that step. It also supports multi-agent coordination and reusable team templates. That structure is what makes it a better fit for teams that need fine-grained control and batch production.

Core capabilities

AI scriptwriting — Driven by leading large language models with strong context handling throughout. Feed in a creative direction or upload a reference script and it generates a structured screenplay with character profiles, scene descriptions and episodic plots. Multiple paths are supported: comic dramas (urban, historical, suspense, romance and other mainstream genres) and commercial films (product selling-point breakdowns, brand advertising scripts).

AI storyboarding — Automatically breaks the script into shot-by-shot boards, planning composition, character blocking, camera movement and transitions, then outputs a structured storyboard sheet including three-view character sheets and multi-camera setups. Storyboard logic is optimized separately for comic dramas and commercial films.

Node-linked image generation — Several leading image models are integrated. Reference images are wired directly to generation nodes, so character likeness and scene style carry through precisely, keeping visual consistency of characters, sets and props across a whole comic-drama season or an entire commercial film. Any node can be re-run in isolation without affecting the others.

Node-linked video generation — Several leading video models are integrated. Storyboards or stills are turned into high-quality, coherent footage in styles ranging from comic drama to human-like drama, with commercial films routed to whichever model best suits the product scene.

Video post-production — Built-in upscaling, video stitching and editing let you finish assets inside the canvas instead of exporting to third-party tools.

AI voiceover and subtitles — Built-in voice generation and subtitle creation support multiple languages, producing vertical HD masters ready to publish to major short-video platforms.

Canvas utilities — Everyday essentials such as image expansion, one-tap 3×3 grid cropping and image trimming, covering the scattered small needs across production.

Team canvas — Multiple people share one canvas, with assets, versions and revision history visible in one place, so a new member can open the canvas and grasp the whole project.

Memory accumulation — Brand tone, visual guidelines and past projects settle into long-term team memory that agents can call on directly, so hard-won experience does not walk out the door with staff.

Batch production line — Once a workflow works, save it as a template; swap the product, selling point or platform ratio and release batches, with vertical, horizontal and square versions switching in one click.

How to use it

Open the platform — Visit the VibePaper official site, register, enter the canvas, and start from a blank canvas or a template.

Describe the creative goal — The screenwriting agent picks a suitable language model, breaks down the brief and produces a structured script.

Generate storyboards — The storyboard agent splits the script into a shot sheet automatically; rework individual nodes you dislike and keep the rest.

Generate and refine visuals — Wire reference images into generation nodes; the visual agent renders frames per storyboard, and any single shot can be fine-tuned independently without disturbing other nodes.

Finish the film — The video agent turns images or storyboards into continuous footage, then built-in post tools handle editing, upscaling, voiceover and subtitles.

Save a template — Store the working flow as a template for the team to reuse, or swap in new content to output batches.

Pricing

The platform is billed monthly and also supports memberships, credit packs and enterprise services.

• Basic — ¥79 per month, 490 credits monthly, roughly up to 245 images or 61 videos, generation concurrency x4, Paper Agent unlocked.
• Pro — ¥199 per month, 1,260 credits monthly, roughly up to 630 images or 157 videos, generation concurrency x8.
• Pro+ — ¥419 per month, 2,800 credits monthly, roughly up to 1,400 images or 350 videos, unlimited concurrency.
• Ultra — ¥819 per month, 5,600 credits monthly, roughly up to 2,800 images or 700 videos, unlimited concurrency.
• Enterprise plans are customizable; contact the team for a quote.

All tiers cover the same selection of language, image and video models; the differences lie in concurrency, credit allowance and monthly fee.

Why it stands out

• Node-level rework. Change one shot and only that node reruns; polish repeatedly without gambling the whole film.
• Consistency by wiring. Reference images connect directly to generation nodes, locking character likeness and scene style so visuals never drift across a season or a campaign.
• A divided AI team. Planning, screenwriting, visual and editing agents each handle their own job, driven by top language models, with a transparent process you can intervene in at any moment.
• Team memory. Brand standards, production know-how and past projects stay in the canvas, so newcomers ramp up fast and a growing team never starts from zero.
• Everything on one canvas. From script to master, images, video, post, voiceover and subtitles all finish inside a single canvas — no tool switching, no lost context.

Where it fits

• Short-drama studios — Several people divide the work while the four agents advance in parallel alongside each role; style standards settle into team memory, serialized content is produced in batches, and newcomers get up to speed quickly.
• Professional short-drama creators — The four agents fill the execution-side manpower while creators keep control of every key decision, completing comic dramas and human-like dramas end to end on one canvas.
• Commercial advertising teams — Brand visual assets are wired in and reused, product selling points are broken down into scripts, and multiple versions are adapted per platform in batch.

Platform support and key specifications

The product takes the shape of a node-based infinite canvas plus four collaborating agents. The planning, screenwriting, visual and editing agents are powered by language models such as the Gemini, GPT and Claude families alongside video models such as Seedance, Kling and Veo. Output covers the major short-video platforms — vertical, horizontal and square versions switch in one click, removing the usual rework for ratio adaptation. Billing is monthly subscription, with membership plans, credit purchases and enterprise services also available.

FAQ

A common worry: if you modify a shot in the middle, does everything before it get thrown away? Under the node architecture, every element — script, storyboard, individual shot, voiceover track — is an independent node. Adjust the shot at second 51 and the nodes covering the first 50 seconds stay exactly as they were; the agent regenerates only that node, leaving the overall chain intact.

As for model choice, the platform integrates multiple leading video generation models, and the visual agent calls the most suitable one for the current storyboard scene. Because the list keeps changing, check the official site for the current lineup. On collaboration, several people can work on the same canvas in real time with assets, versions and decisions managed centrally; for team permissions and billing details, refer to the official pricing page.