Yarn - Speedrun verification tool
​A professional tool that eliminated a decade-old redundant workflow in speedrun verification through open-source data auditing.
​
-
Solo project
-
React / TypeScript / Tailwind CSS​
-
Web-based software development
Try Site: https://kr0ss0.github.io/yarn/
Marking a load and Verifying loads by skipping to each marked frame.
GitHub Page
Yarn is a professional-grade tool designed to modernize the way competitive speedruns are verified. Speedrunning is the act of completing a video game as fast as possible, often requiring frame-perfect execution. Because players use different hardware, many communities use Load Removed Time to ensure that a player with a faster computer/console does not have an unfair advantage over someone with a slower computer/console.
For over a decade, this process was fundamentally inefficient. Runners would time their own loads, only for moderators to re-time the entire video from scratch to prove the data was accurate. Yarn eliminates this redundant labor by turning timing into an auditable data format. It allows runners to export their markers as a JSON file so that verifiers can simply audit the existing data in seconds rather than up to an hour.
​
Precision
Standard web video players are not precise enough for competitive timing. For Yarn I implemented a custom wrapper for the YouTube API that enables frame-by-frame navigation with millisecond precision. This engineering ensures that every timing marker is accurate to the exact frame of the video source.
​
Technical Implementation
The application is built with React and TypeScript to handle complex state management and ensure type safety across the timing engine. It utilizes a zero-database architecture, relying on persistent local storage hooks and file-based portability to ensure user data remains private, secure, and instantly shareable.
​
Transparent Data Standards
Yarn moves the community toward an open-source model of verification. By standardizing timing data into portable JSON files, the "black box" of moderation is removed. Any member of the community can import a Yarn file to see exactly which frames were used for a World Record, ensuring total transparency and competitive integrity across the board.
​
The Audit Workflow
The core innovation of Yarn is the Verifier Audit Cycle. Instead of manually scrubbing through a video, a moderator can use automated jump-points to snap the playhead to the exact frame before and after a load screen. This transforms the verification process from a data-entry task into a rapid visual confirmation, reducing the total administrative workload by over 95%.
​
Try Site: https://kr0ss0.github.io/yarn/
GitHub Page
Validation
Yarn includes a real-time validation engine designed to eliminate human error during the timing process. The system automatically monitors the marker timeline to detect logical inconsistencies, such as overlapping load screens or markers placed outside the boundaries of the run. By instantly flagging invalid durations and conflicting timestamps, the tool ensures that every exported dataset is mathematically sound before it ever reaches a moderator.
