AI Video Generation Tool (Batch Supported)

sor2 model, supports text-to-video and image-to-video, can generate AI videos individually or in batches, and offers multiple model options and parameter settings.

Bug report

Loading tool, please wait...

If the online tool fails to load or run, try the desktop tool.https://tools.yikeaigc.com/

Tool Usage

Return to old version

API Settings

API Domain

Enter the domain provided by your API provider

API Key

Model Selection

1Enter prompt

Describe the video scene, actions, style, etc. More detail gives better results.

2Reference image (optional)

Upload a reference image to better match the desired style. Supports JPG, PNG, etc.

3Generation Settings

Quantity

Generate multiple videos with the same prompt (1-10)

Example Prompts

Landscape

At sunset, golden sunlight shines on a calm lake, distant mountains appear faintly, and a white egret flies slowly by

Portrait

An elegant woman strolls under cherry blossom trees, petals drifting in the wind. She wears a light-colored dress and smiles

Sci-Fi

A night view of a futuristic city, neon lights flashing, flying cars weaving between skyscrapers, raindrops falling on holographic billboards

1Batch prompt input

Text input

Upload TXT Files

One prompt per line. Videos will be generated in order.

2Reference image (optional, shared by all videos)

All batch-generated videos will use the same reference image

3Batch generation settings

Concurrency

Concurrent tasks (1-5, 2-3 recommended)

Failed Retry Count

Retries on failure (0-3)

Instructions

Software Usage Instructions

Configure API Settings:
- API Domain: Enter the API domain address provided by your AI video service provider
- API Key: Enter the API key (Key) you obtained
- Model Selection: Choose portrait (9:16), landscape (16:9), or a custom model
Select Generation Mode:
- Single Generation: Suitable for testing or generating a small number of videos. You can use the same prompt to generate multiple videos
- Batch Generation: Suitable for producing large numbers of videos. Supports text input or uploading a TXT file to import multiple prompts
Enter Prompts:
- Single mode: Enter a detailed video description in the text box
- Batch mode: Enter one prompt per line, or upload a UTF-8 encoded TXT file
- Click example prompts to quickly fill in reference content
Upload Reference Image (Optional): Uploading a reference image can make the generated video closer to the expected style. JPG, PNG, and other formats are supported. In batch mode, all videos will share the same reference image.
Set Generation Parameters:
- Generation Quantity: In single mode, you can set the number of videos to generate repeatedly (1-10)
- Concurrency: Number of tasks processed simultaneously in batch mode (recommended: 2-3)
- Retry on Failure: Number of automatic retries when generation fails (0-3)
Start Generation: Click the "Start Generating Video" button, and the system will call the API to begin generation. You can view real-time status in the progress bar and log area, and click "Stop Generation" to interrupt the task if needed.
Download Results: After generation is complete, you can preview the first 20 videos. Click "Download All Results (ZIP)" to package and download all generated video files in one click.

FAQ

A: You need to register an account on an AI video service provider platform and apply for an API key. The application process may vary slightly between providers. You can usually create and manage keys on the platform's "API Management" or "Developer Center" page. Please keep your key secure and do not disclose it to others.

A: Text-to-Video (T2V) generates a video based only on a text description, with the AI independently creating the video content; Image-to-Video (I2V) provides an additional reference image on top of the text description, and the AI uses the image’s style, composition, or elements as a reference to generate the video, making the result more controllable.

A: The portrait model (9:16) is suitable for creating content for short-video platforms such as Douyin, Kuaishou, Xiaohongshu, etc.; the landscape model (16:9) is suitable for creating content for platforms such as Channels, Bilibili, YouTube, etc. Please choose the appropriate video aspect ratio based on your publishing platform.

A: A good prompt should include the following elements: 1) subject description (person, animal, object, etc.); 2) action or state; 3) scene environment; 4) lighting and atmosphere; 5) visual style (realistic, animated, cinematic, etc.). The more detailed and specific the description, the closer the generated result will be to your expectations. We recommend testing with sample prompts first, then creating your own content once you are familiar with it.

A: We recommend setting it to 2-3. Too high a concurrency value may cause API rate limiting or an increased generation failure rate, while too low a value will reduce efficiency. If generation fails frequently, try lowering the concurrency. We also recommend setting 1-2 retry attempts for failures to improve the overall success rate.

A: Video generation time depends on several factors: the API provider’s processing speed, prompt complexity, whether reference images are used, and more. Typically, generating a single video takes anywhere from 30 seconds to 3 minutes. During batch generation, the tool automatically manages the task queue, and you can view the status of each task in the log area.

A: Possible reasons for generation failure include: 1) the API key is invalid or has insufficient balance; 2) the API domain is entered incorrectly; 3) the network connection is unstable; 4) the prompt contains sensitive content and is blocked; 5) a temporary issue with the provider’s API. Please check the error message in the logs, confirm that the API configuration is correct, and contact the provider’s customer support if necessary. The tool has a built-in automatic retry mechanism, and some temporary errors will recover automatically.