This repository is a sample generative AI video short-form generator application that uses Amazon Bedrock, Amazon Transcribe orchestrated with AWS serverless services. This sample application takes in long videos, understands it, and extracts and generate up to 15 short-form highlight video. It uses distant parts from the video to make more than 10 engaging, short-form content with titles and subtitles. The app also gives flexibility in choosing which frame to put in, and also the ability to edit titles and subtitles auto-generated by LLMs.
aws-shorts-demo.mp4
-
Install AWS CLI & Set up AWS credentials
Important
- The application uses Amazon Bedrock in the us-west-2 region. Please allow model access in us-west-2.
- The application only supports models from Anthropic Claude 3.0 and above (3.0 Haiku, 3.0 Sonnet, 3.0 Opus, 3.5 Sonnet).
- Revised and Improved Prompt for extracting and processing topics
- Added feature to edit video in a more vertical ratio (3:4)
- Improved Timeframe Matching accuracy after highlight extraction
- Added "Upload with Subtitles" Option for faster processing
- Added Claude 3.5 v2 as an option
- Clone repository
git clone https://github.com/aws-samples/gen-ai-video-short-form-generator.git
- Install dependencies
cd gen-ai-video-short-form-generator
npm install
- Deploy cloud sandbox
npx ampx sandbox
Important
- It takes about 10 minutes to deploy.
- Do not terminate the sandbox environment while running the front-end application.
- Run frontend app
npm run dev
Create your own repository and follow Amplify deployment steps
npx ampx sandbox delete
Important
You can verify whether all the resources have been deleted in the AWS CloudFormation console.
To delete an Amplify project that has been deployed from the Amplify Development Step,
- Go to your Amplify project console
- Navigate to
App Settings > General Settings > Delete app
See CONTRIBUTING for more information.
This library is licensed under the MIT-0 License. See the LICENSE file.
For detailed information about the open source libraries used in this application, please refer to the ATTRIBUTION file.