Automated Banner Image Generation for Workflow Prompts for AI-Labs
Summary
Automated the generation and injection of banner images for all Markdown prompt files in
/content/lost-in-public/prompts/workflow
using a custom Python script and the Recraft API.Why Care
This workflow ensures every prompt has a visually relevant, AI-generated banner image, improving the visual quality and consistency of documentation. It also enforces security and maintainability by using environment variables and externalized style configuration.
Implementation
Changes Made
- Set up Python virtual environment in the
a-labs
submodule. - Created and configured the Recraft API key and other parameters necessary for successful AI image generation.
- Selected five sample images that allowed us to develop our own "style" for the Recraft API.
- Updated and ran
ai-labs/recraft/generate-banner-images-recraft.py
to process all prompts for the prompt library. - Script loads a custom style from a JSON file and securely loads the Recraft API token from environment variables.
- Robust error handling for missing tokens or invalid styles.
Technical Details
- The script uses the
image_prompt
field to generate images via the Recraft API and injects the returned URL into thebanner_image
field. - All API tokens are loaded from environment variables; no sensitive data is hardcoded.
- Custom style configuration is stored in a JSON file for reproducibility.
- Aggressive commenting and DRY principles followed throughout the script.
Integration Points
- All prompt files in
/content/lost-in-public/prompts/
now include abanner_image
field. - No breaking changes to existing metadata or content structure.
- Future prompt/image generation workflows should build on this pattern.
Documentation
- See
/ai-labs/recraft/generate-banner-images-recraft.py
for script logic and usage. - used the Ask Generative AI model to generate a Style to get GPT 4.1 to generate the image_prompt values.
- All changes follow the changelog entry guidelines in
/content/lost-in-public/prompts/workflow/Write-a-Code-Changelog-Entry.md
. - The script workflow is now fully recursive and supports robust error handling for missing tokens or invalid styles.
Content & Metadata Enhancements
- YAML Frontmatter Updates:
- Added or updated
banner_image
fields in all processed workflow prompt files, ensuring each has a visually relevant, AI-generated image. - Maintained strict preservation of all existing metadata and formatting.
Security & Configuration
- Environment Variables:
- All API tokens are securely loaded from environment variables; no sensitive data is hardcoded.
- Custom style configuration is externalized to a JSON file for easy updates and reproducibility.
Code Quality & Documentation
- Aggressive Commenting and DRY Principles:
- Script and workflow changes are thoroughly commented, following the project’s aggressive documentation standards.
- Modular, single-source-of-truth approach for style and token management.
Next Steps
- Review generated banner images for creative quality and relevance.
- Continue to iterate on prompt and image generation workflows as needed.