AI Art Engine 中文
Menu

AIArtEngine · Short drama · Ads · Film

Professional AI creation tool

v5.0: A built-in MCP tool server lets Claude Code and other AI agents drive the app; the AI chat panel @-references assets and calls generation tools right in conversation; One-Click Workflow covers 12+ industry templates. Assets · shots · node graph in a single orchestration flow — projects stay on your disk, models run on your own keys.

Node graph Generation chains, asset preview and task execution
Episode pipeline Beat breakdown, 9-grid / 4-grid storyboards and director review
3D director stage Blocking shots, motion recording and camera views
Multi-angle Fast camera and framing switches
Lighting Fine control of light and atmosphere
Mood Control the emotion and color grade of a shot
Image edit Local repaint and image revision

WindowsmacOSLinuxGPL-3.0

A professional pipeline, with your files still on your disk.

From script and shots to node generation, the director stage and the final timeline — short drama, ads and narrative films all share one workflow; projects are local-first and never forced into the cloud. The short-drama Agent pipeline generates a complete node graph in one click, then handles 9-grid, 4-grid, director review and motion video inside a single pipeline window. A built-in MCP tool server means the in-app AI chat panel can @-reference assets and generate directly, and external agents like Claude Code can plan workflows, run generation and edit node graphs in conversation.

AI chat + MCP, the headline features

AI here does more than chat — it gets work done: drive generation and workflows right from an in-app conversation, or let external agents like Claude Code connect to your project directly.

In-app AI assistant

AI chat panel

@-reference project assets in conversation and let the agent call MCP tools directly — generate image / video / 3D / speech, edit node graphs, run workflows and track status. Multi-session history; pick any configured text provider for the model.

Manual §5.5 →
Driven by external agents

MCP tool server

Built-in MCP Server: external AI agents such as Claude Code / Codex can drive your project over a stdio bridge or direct HTTP — plan & commit workflows, run generation, read/write assets and node graphs. Stable tokens across restarts, audit log, concurrency gate.

MCP setup guide →

From node wiring to a finished cut

One generation chain: reference image → character design sheet → select image → video generation.

Node graph example: reference an image to generate a character sheet, then select images to generate video
Node wiring
Video output
  1. 01

    One-Click Workflow

    Pick a preset (including the “short-drama storyboard” short-video template) or describe what you need in one sentence, preview the topology and create a reusable host asset; drag it onto the canvas to wire it up and run. See Manual §7 and the Short-Video Guide.

  2. 02

    Asset library

    Images, videos and voices are referenced by GUID, and asset packages can be imported and exported.

  3. 03

    Episode pipeline

    Beat breakdown, 9-grid / 4-grid storyboard tables and director review turn narrative pacing into shots.

  4. 04

    Node generation

    Text / image / video / voice / 3D model nodes; works with OpenRouter, OpenAI, Google, Volcengine Ark, Kling, MiniMax, Tongyi Qianwen, ModelScope, ComfyUI, Meshy, Tripo, Rodin (Hyper3D), Luma AI and Lux3D, plus configurable multi-cloud object storage. Port types must match, and select nodes only accept list ports.

  5. 05

    Director stage

    3D posing and cameras; supports a 3D model input port; capture blocking shots and record motion clips, then feed them downstream through the square ports.

  6. 06

    Final timeline

    Place imported and node-produced media on tracks, preview the selected clip or play the whole timeline, then export the cut.

  7. 07

    Agent pipeline

    The short-drama template auto-generates script → beats → 9-grid → tiling → 4-grid → motion prompts → 36 videos; drive everything from the Episode Pipeline window, where regenerating a stage invalidates downstream stages and re-runs them consistently.

Get the installer

Pushing a v* tag lets GitHub Actions build multi-platform artifacts automatically. On Intel Macs download the dmg whose filename contains x64; on Apple Silicon (M series) pick arm64.

Community & feedback

Join the QQ group to discuss usage, watch the video tutorials, or reach out by email.

AIArtEngine QQ group QR code, group number 647306826
Scan to join · Group 647306826