AI Comic Show Generator
The AI-driven comic drama generator supports fully automatic pipelines such as script creation, character extraction, intelligent lens splitting, and video generation.
##Function overview | function| description| |------|------| | 🎬 AI script generation| Enter creative ideas and AI automatically generates a complete script| | 👤 Automatic role extraction| Intelligently identify characters from scripts and generate four-view reference maps| | 🎞️ Intelligent splitter| AI automatically splits scenes, generates split mirror descriptions and action scripts| | 🖼️ First and last frame generation| Generate key pictures of the first and last frames for each mirror| | 🎥 video generation| Generate animated video clips based on key frames or reference pictures| | 🎬 video synthesis| Composes all clips into the final video and supports subtitle burning| | 🌐 multilingual| Chinese/ English /Japanese/ 한국어| ##Start quickly ###Step 1: Configure the AI model (must be done for the first time) Click on the ** ˇ Settings ** icon in the upper right corner to enter the model configuration page.️You need to configure three types of models: #### 1. Text model (used for script generation and lens description) | agreement| recommendation model| Base URL | |------|---------|----------| | OpenAI |`gpt-4o`,`deepseek-chat`, etc.|`api.openai.com`(or compatible API address)| | Gemini |`gemini-2.5-flash` and so on| `https://generativelanguage.googleapis.com`| #### 2. Image model (used for character reference pictures, end and end frames) | agreement| recommendation model| Base URL | |------|---------|----------| | OpenAI |`gpt-image-1` and so on| `https://api.openai.com`| | Gemini |`image-3.0-generate-002` and others| `https://generativelanguage.googleapis.com`| | Kling |Creable image model| `api.klingai.com`(requires AK/SK)| #### 3. Video model (used for video clip generation) | agreement| recommendation model| Base URL | |------|---------|----------| | Gemini (Veo) | `veo-2.0-generate-001` | `https://generativelanguage.googleapis.com` | | Seedance |ByteDance is Dream Video Model| `https://ark.cn-beijing.volces.com`| | Kling |Credible video model| `api.klingai.com`(requires AK/SK)| ** Configuration steps: ** 1. Go to the ** Settings ** page 2. Click ** Add supplier ** 3. Fill in the supplier name (e.g."Gemini") 4. Select agreement type 5. Fill in the Base URL and API Key 6. Click ** Get Model ** to automatically pull the model list, or manually enter the model ID 7. Check the model to use 8. Select one default model for "Text/Image/Video" in the ** Default Model ** area --- ###Step 2: Create a project 1. On the home page (My Projects), click ** New Project ** 2. Enter the project title and click Create --- ###Step 3: Four-step workflow After creating the project, the left navigation shows four steps: #### Step 1ˇ Script 1. Enter your story idea in the ** Creative Ideas ** text box (for example: "A lazy cat, likes to program, writes code with glasses") 2. Select the text model (top drop-down) 3. Click **AI to generate script ** → AI will stream and output the complete script 4. The script content can be manually edited after generation 5. Click ** Save ** #### Step 2ˇ Role 1. Click ** Character Analysis ** → AI automatically extracts all characters from the script 2. Each role card displays a name and description, which can be edited 3. Select image model 4. Click the ** of a single character to generate four views ** or use ** to generate four views in batches ** 5. AI will generate a reference map for each character that includes front, side, back, and 3/4 angles 6. click on the image to enlarge #### Step 3ˇ Split the mirror The mirror split page is the core operation area and supports two generation modes: ** Mode A: Based on the end and end frames (Keyframe mode)** 1. Click ** Generate Split Lens ** → AI splits the script into multiple shots 2. Each lens segment contains: scene description, first frame description, last frame description, action script, lens direction, duration, and lines 3. Click ** to generate first and last frames in batches ** → generate first and last frame pictures for all mirror segments 4. Select video model and picture scale (16:9 / 9:16 / 1:1 / Auto) 5. Click ** Batch Generate Video ** → Generate animated videos based on the first and last frames ** Mode B: Based on reference diagram (Reference mode)** 1. Switch to the ** Based on Reference Picture ** tab 2. AI uses character reference map + scene description to directly generate video 3. No need for first and last frame steps, the process is faster ** Single mirror split operation: ** - Expand the mirror split card (click on the ►) to view and edit the detailed description - ** Regenerate text ** -Ask AI to rewrite the description of the mirror - ** Generate picture ** -Generate the first and last frames of the mirror separately - ** Generate video ** -Generate video for this split lens separately - ** Copy prompt words ** -Copy full prompt for generating video - Adjustment duration (5-15 seconds) #### Step 4ˇ Preview 1. After all video clips are generated, 2. Click ** Compose Final Video ** → Automatically splice all clips and burn subtitles 3. Click ** to download the video ** to get the final product 4. Click ** Download Material ** to download material packages for all first frames, last frames, and video clips ##ˇFrequently Asked Questions | problem| solutions| |------|---------| | Prompt "Please configure the model first"| Go to the settings page to configure the corresponding type of AI model| | generation failed| Check whether the API Key is correct and whether the model is available| | Video generation is slow| This is normal, video generation usually takes 1-5 minutes per video| | Unable to generate videos in batches| The first and last frames need to be generated first (in Keyframe mode)|
