ArchitectAI leverages the power of the Large Language Model Gemini to analyze software architectures. By sending parallel requests to the LLM and storing the results in a database, it optimizes response speed and efficiency. The user-friendly interface provides multiple features that allow users to capture, upload, and analyze images or text transcriptions of software architectures.
Home Page:
- Dark Mode
- Help Button: Access a guide to using the app.
- History: Save analysis results for future reference.
Main Functionalities:
- Capture Image: Use the CameraX feature to capture high-quality images of software architectures.
- Upload Image: Upload an image from your device.
- Upload Text Description / Audio Transcription: Submit a text-based input of a software architecture for analysis.
Image Handling in camera and preview page:
- Zoom: users can zoom while taking a picture to avoid impractical parts.
- Edit: After capturing or uploading an image, users can crop the image and edit it.
- Additional Context: swipe the page to add additional textual information about the architecture before submitting.
Analysis Feature page:
- Request Validation: Ensures the submission is a valid software architecture.
- LLM Analysis: Submits the image and additional information to LLM Gemini for analysis, displaying the result in a formatted HTML view.
- Output Customization: Choose from various outputs like Component Diagram or State Diagram, and view the architecture visualized using PlantUML/C4 format.
- Code Access: View and copy the PlantUML code for the generated diagrams.
- Enhanced Architecture: Option to get an improved version of the architecture based on user-defined filters such as energy efficiency, sustainability, security, and maintainability.
- History: Allow the user to choose whether to save the result or discard it.
Additional Features:
- Summary Button: Summarizes long text to improve readability, can then copy it for further usage.
- Help Section: A detailed guide with an explanatory video on how to use the application.
- Sharing and Saving: Export results as PDFs or save images for future use.
The demo video can be viewed here
- Download the Source Code
- Configure API key
- Get an API key for Gemini at https://aistudio.google.com/apikey
- In your local properties file, type
apiKey = <your key>
- Start the App
- Capture or Upload an Image/Text
- Edit and Preview: Use the provided tools to zoom, crop, or add additional text before submitting.
- Analyze: Submit your image or text for analysis. Review the results and choose from various output options to visualize your architecture.
- (optional) Save and Share: Save your analysis to the history or export the results as a PDF for sharing.
We welcome all contributions and improvements ideas! If you have suggestions or improvements, please open an issue and discuss your ideas or fork the repository and submit a pull request.
This project is licensed under the MIT License. See the LICENSE file for details.