01
Human authority
Marks and released feedback remain the responsibility of educators—not an automated system.
A new chapter
Assessment needs more than an answer.
Feedbacker is becoming a workflow, governance, and assessment layer around AI—helping educators work responsibly, consistently, and at scale.
The project is being redefined and rebuilt in the open.
What endures
AI assists.
Educators decide.
02
Visible provenance
Inputs, criteria, transformations, suggestions, edits, and approvals should be traceable.
03
Institutional control
Models, prompts, data handling, retention, and deployment should be governable and replaceable.
04
Cohort consistency
Rubric-led workflows should support moderation and calibration without flattening academic judgement.
The direction
Responsible assessment infrastructure, built with educators.
The previous application has been retired. Its successor starts with governance, privacy, accessibility, and the realities of academic practice—not with a model or a prompt.
Follow the project on GitHub