Accounts, Plans & Credits

Overview

AI model for all evaluations

Pins ONE Claude model for every evaluation by every teacher (Evaluate Now and Economy Batch). Teachers are not told, and have no model choice of their own: their screens never name the model. It applies to the next request sent; evaluations already running finish on the model they started with. Choose “Not pinned” to return to the default, Sonnet 5.5.

Evaluation mode for all teachers

Decides which evaluation modes EVERY teacher is offered. With Both, teachers see Evaluate Now and Economy Batch and choose. With a single mode, the teacher's AI Evaluation section is replaced by the message "Press the Start Evaluation button and the result will be available after some time", and the server refuses the other mode. Runs already in progress carry on. Takes effect for teachers within seconds.

Evaluation failures

Technical failures only (the AI service busy or unreachable, an Economy item that failed, an AI reply that could not be read). Not marks or academic judgments. No student names are kept.

Answer sheets evaluated, by teacher

How many answer sheets each teacher had evaluated in a period, and how many attempts failed. Days are Indian Standard Time days, both dates included. Evaluated = a saved result (Evaluate Now or Economy Batch), counted on the day it finished. Failed attempts = technical failures from the failure log; a sheet that failed and then worked on retry appears in both columns. Counts only: no student names or answers.

Manage Plans

Plans are a global catalog (not institute- or user-scoped). Editing a plan here only ever affects FUTURE purchases — every existing order/payment/credit already captured its own price/credit snapshot at the time it was created and is never recalculated.

Accounts

Account Detail