Skip to main content
ClassTrack turns a recording of a real lesson into written feedback for the teacher who taught it. You send the recording, the class details, and which teaching rubric to use, and ClassTrack transcribes the lesson, judges it against each part of the rubric, and quotes the lesson back as evidence. A rubric is a list of criteria you judge teaching against, each with a description and a score range. Everything comes back as writing, including the scores. There are no numbers to chart.

​
What does the teacher submit?

One recording per analysis, as a link, plus the details of the class. Everything in this table is required except the institution ID, so ClassTrack needs more than a recording and a rubric to run.
The recording is not uploaded through ClassTrack. You put the MP3 somewhere it can be downloaded, such as your own cloud storage, and send the link.

​
What does the analysis contain?

Five parts, all in writing:

An overall level

One overall judgment of how effective the teaching was, written as free text.

A summary

A few paragraphs on the lesson: what worked, what to improve, and how the class went.

A judgment per rubric criterion

For every criterion in your rubric: a score, what was observed, what to do about it, and examples taken from the lesson. The score here is free text too.

Classroom insights

Observations, improvement strategies, and examples across whichever areas the AI decided to comment on.
Both scores are words, not numbers. The overall level and each criterion’s score come back as free text written by the AI, with no scale, no percentage, and no star rating. If your product needs a number to store, compare, or chart, work it out on your side from the text, and expect the wording to vary between lessons.
The AI is asked to cover instructional clarity, student engagement, classroom management, differentiation and support, assessment and feedback, and the teacher’s strengths. That list is a prompt rather than a fixed structure, so the areas it names, and how many come back, can differ from one analysis to the next. Do not build screens that expect exactly those six. The fifth part is an action plan, split into three time frames: You also get the transcript of the lesson, so anyone reading the feedback can check it against what was actually said.

​
Who reads the result?

The teacher who was recorded, and whoever coaches them: a mentor, a head of department, or a principal. The feedback is written as coaching, addressed to the teacher, with evidence quoted from the lesson and concrete next steps. Because each analysis carries your teacher ID and an optional institution ID, your platform can group a teacher’s analyses over time.

​
What are the limits?

Audio only

MP3 recordings. Video is not supported.

One rubric per analysis

Each recording is scored against one rubric.

English and Hindi

The transcription recognizes English and Hindi.

Finished lessons

A whole recording is analyzed after the lesson. There is no live analysis while teaching.
ClassTrack works from what was said in the room. It does not identify who is speaking, so it does not report how much of the lesson the teacher talked compared with the students.

​
How do I add ClassTrack to my platform?

Your developers connect to it through the ClassTrack API.

How it works

What happens between sending a recording and reading the feedback.

Developer guide

Analyze a recording with the API, step by step.

​
Frequently asked questions

No. ClassTrack accepts MP3 audio only. If you recorded video, take the audio track out of it and send that.
Yes, and you should. ClassTrack scores against the rubric you name, and rubrics and their criteria are yours to create and edit through the Rubrics API.
No. The lesson is transcribed as one flow of speech, without separating the teacher from individual students. Observations about engagement are based on what was said and asked in the room, not on measured speaking time.
Ask for the analysis at any time and it tells you whether it is still in progress. If your platform has a webhook set up, which is an automatic message CrazyGoldFish sends your server, you are told as soon as an analysis finishes or fails.
Not in ClassTrack. The analysis is returned to your platform as it is, and your platform decides who sees it and when.

​
Next steps

Use cases

How schools and platforms use ClassTrack.

Developer guide

Analyze a recording with the API, step by step.