Enhanced image handling in CorrectionModal and RegenerationModal components to utilize fallback URLs for improved reliability. Updated auto insurance strategies for better alignment with ad formats and removed unused visual guidance. Added GLP-1 image creative prompts for more effective ad generation.
- Updated API endpoints to include optional parameters for trending topics in ad generation requests.
- Added new endpoint for bulk exporting ads as a ZIP file containing images and an Excel sheet.
- Implemented trend monitoring service to fetch relevant news articles for specified niches.
- Enhanced validation schemas to accommodate new parameters for trending topics.
- Updated generator service to incorporate trending context into ad copy generation.
- Created export service for handling bulk ad exports, including image downloads and Excel file creation.
- Added caching mechanism for trending topics to optimize performance.
feat: add cancel button to generation progress and fix stuck loading state
feat: Enhance AI image generation prompts for natural, photorealistic advertising creatives by enriching context and refining modification instructions.
- Added a preview_only option to the image regeneration API, allowing users to generate images without immediate database updates.
- Implemented a new ImageSelectionRequest model for confirming user selection between original and regenerated images.
- Updated the frontend RegenerationModal to include a comparison step, enabling users to choose their preferred image before saving.
- Enhanced API endpoints and types to support the new selection confirmation feature, ensuring a seamless user experience.
Implement image regeneration feature and update API endpoints
- Added a new API endpoint for regenerating images, allowing users to create new versions of existing ads with optional model selection.
- Introduced a frontend modal for image regeneration, enhancing user interaction and experience.
- Updated the API to list available image generation models, providing users with options for customization.
- Refactored data models and types to support the new regeneration functionality, ensuring type safety and consistency across the application.
- Enhanced the concepts framework by moving visual structure layouts to a separate file for better organization and clarity.
Refactor niche type declaration in form components
- Updated the niche property in BatchForm, ExtensiveForm, and GenerationForm components to use 'as const' for improved type safety and consistency in TypeScript.
Enhance extensive ad generation with custom niche support
- Updated the ExtensiveGenerateRequest model to include an optional custom_niche field for enhanced flexibility in ad generation.
- Modified the generate_extensive function to validate custom niche input when 'others' is selected as the niche.
- Adjusted frontend components to accommodate the new custom_niche field, including validation in the ExtensiveForm and updates to the GeneratePage.
- Updated API endpoints to handle the new custom_niche parameter, ensuring seamless integration across the application.
Refactor extensive ad generation to support batch responses
- Updated the API endpoint to return all generated ads instead of just the first one, enhancing the batch generation feature.
- Modified the frontend components to handle and display batch results, including updates to the GeneratePage and ExtensiveForm for better user feedback.
- Changed variable names and comments for clarity, ensuring consistency in terminology related to ad images and variations.
- Improved error handling and logging in the generator service to accommodate the new batch processing logic.
Update ad generation features and enhance user input options
- Renamed project references from Ad Generator Lite to PsyAdGenesis, reflecting the new branding.
- Added optional fields for target audience and offer in ad generation requests to improve customization.
- Enhanced the GeneratePage and BatchForm components to include input fields for target audience and offer.
- Updated API endpoints to handle new parameters for ad generation and batch processing.
- Improved error handling and validation for new input fields to ensure robust user experience.
- Enhanced the ad creative generation logic to incorporate user-defined target audience and offer in the output.
Rename project from Ad Generator Lite to PsyAdGenesis, updating all relevant references and configurations. Enhance descriptions and metadata to reflect the new branding and focus on ad design. Implement sorting functionality in the gallery for improved ad management.
- Added verification and logging for generated images in GeneratePage and AdPreview components to improve user feedback and debugging.
- Updated API endpoint to ensure required parameters are always sent for image generation.
- Enhanced image generation logic in the generator service to support parallel processing and improved error handling.
Update frontend dependencies and enhance UI components for better user experience
- Updated Tailwind CSS version to 4.1.18 in package.json and package-lock.json.
- Introduced a reusable GradientBadge component for consistent styling across the application.
- Improved layout and typography in DashboardAdCard and AdCard components for better readability and interaction.
- Added loading states and error handling in new AnglesPage and ConceptsPage components.
- Enhanced progress components in the generation workflow to provide better user feedback during ad generation.
Initial commit of the Ad Generator Lite project, including backend services, frontend components, and configuration files. Added core functionalities for ad generation, user management, and image processing, along with a structured matrix system for ad testing.
