#Autoscroll Recorder

• Added "unknown" status detection for jobs that get lost in the pipeline, making it easier to spot and retry failed jobs.
• Improved job tracking by switching from job IDs to SQS message IDs for more reliable ECS task matching.
• Added batch job tracking and repository cleanup features.
• Built a reusable jobs table component for consistent UI across different pages.

• Launched a full API system for programmatic job creation, enabling automated workflows.
• Implemented type-safe API key management with proper authentication.
• Fixed authentication issues with X-API-Key header support for Vercel compatibility.

UX and recording improvements:

• Fixed expired presigned URLs on job details pages.
• Enabled retry for jobs with unknown status.
• Removed unnecessary page reloads after bulk job refresh.
• Improved configuration settings UI with better documentation.
• Hidden the cursor in video recordings using FFmpeg's -draw_mouse option.
• Prevented cookie banners from hiding entire HTML elements.
• Implemented URL-based filename generation with date partitioning for better organization.

#Inspiration Index

• Migrated to a new multi-site scraper architecture using the Firecrawl SDK.
• Launched an Awwwards scraper MVP with daily cron scheduling and backfill capabilities.
• Separated scraping logic from recording workflows for better modularity.
• Added type-safe enums and improved submission handling.
• Implemented proper lock release on script failures with enhanced logging.

• Implemented shared ECR repository infrastructure across environments.
• Updated production configuration for deployment readiness.
• Added ECS task tagging for improved job tracking.
• Fixed ES module compatibility issues for Vercel deployments.