Is Class Companion a good fit for Australian classrooms?

Details about other products were checked on 29 July 2026. Other providers change their features and pricing, so check their site before deciding.

What does Class Companion actually do?

Class Companion is an AI assignment platform. Students answer questions you set inside the tool, and the tool scores their work against a rubric and writes them feedback. Its help centre lists three formats: short answer, essay and multiple choice, plus an ordered outline you build from short-answer questions. Essays can be typed or handwritten, with handwritten pages converted to text by AI so they still receive feedback. You can use built-in rubrics such as SAQ and FRQ, or write your own criteria.

Scoring is rubric-driven rather than free-form, which is the right foundation. Feedback a student can trace back to a criterion they have already seen is feedback they can act on. Running alongside the assignment is an AI tutor that, in the platform’s own words, helps students without giving away the answer. It is on by default for each assignment, you can switch it off, and you can read any student’s full conversation with it.

Subject coverage is broad but shaped by the American system. The help centre says it supports all subjects that have written or multiple choice assignments, including Regular, Honors and AP classes, with templates for English, Social Studies and Science. The built-in rubric types are AP task types: LEQ, DBQ, FRQ and SAQ. The same documentation states the marking runs on OpenAI’s GPT-4.

Who sees the marking first, and can you hold it back?

This is the question that decides whether a tool fits your assessment policy. In Class Companion the feedback is written for the student, but you control the timing. The assignment setup includes a setting for when students receive feedback, instantly or at the end, a limit on how many attempts a student gets per question, and an option to hide the score from students.

You are not locked out of the mark either. A teacher can override any score from the submissions view, and the override asks you to set a final score and add a comment saying why the student earned it. Use that comment field properly rather than clicking past it. It is the only record of why the released mark differs from the one the model produced, and it is exactly what you want in front of you at a moderation meeting or in a conversation with a parent.

Read those settings against your own policy instead of accepting the defaults. For low-stakes practice, instant feedback is the whole point: students revise while the task is still live, and they get more attempts than you could ever mark by hand. For anything feeding a report grade or a moderation meeting, most departments want the mark checked before release, which means using the end-of-task setting or hiding the score and doing a pass yourself. That is a decision to make deliberately, not a box to leave ticked.

What does Class Companion cost?

Its help centre answers this plainly: Class Companion is free for teachers. One teacher can create classes, set assignments and get AI marking without a purchase order. That is genuinely uncommon in this category and worth crediting, and it makes the tool cheap to trial properly rather than judge from a demo video.

The paid school plan is where the school-level features sit. The help centre lists LMS integration, AI and plagiarism detection, standards alignment, a private content library, custom individual learning plans, and class or school progress insights. A separate article describes premium access as adding advanced rubric settings, detailed analytics, two-way grade transfer with Canvas and Schoology, transcription of handwritten essays and priority support. No price is published; teachers are told to introduce their principal or IT admin.

One structural point for anyone evaluating at school level. Class Companion is now part of Panorama Education’s Solara platform, which is sold to districts. If you take it past a single classroom, expect a wider platform conversation rather than a per-teacher subscription.

Does it fit an Australian classroom?

The mechanics travel; the curriculum framing does not. Nothing about rubric-scored short answers and essays is America-specific, and an Australian teacher could get real value out of the free tier this week. But the built-in rubrics are named after AP tasks, and the documentation makes no claim of alignment to ACARA achievement standards or to any state syllabus. Translating your outcomes into custom rubric criteria is your work, not the tool’s.

The integrations lean the same way. Two-way grade transfer is listed for Canvas and Schoology, so check what your school actually runs before you assume marks will flow anywhere. On data, the help centre says it collects the minimum needed to operate, names OpenAI’s GPT-4 as the model, and states that OpenAI does not use its data to train models. The compliance it names is FERPA, COPPA and state student privacy regulations, all American frameworks, so an Australian school still has to run its own privacy assessment rather than treat United States compliance as equivalent.

Be clear too about what it is not. It marks work students submit into it, which does nothing for a student revising alone at home. Those students are still looking for study guides and worked examples somewhere else.

What should you check before putting an AI marking tool in front of a class?

These questions apply to any AI marking tool, not just this one. Work through them before a faculty commits, because unwinding a choice after a term of assessment data costs far more than a careful trial does.

Then add the test people skip. Feature lists are easy to compare and feedback quality is not, so run a real class set through the tool and read ten pieces of what it produced. Class Companion is upfront here: its help centre has an article explaining that identical answers can score differently because language models vary, and it asks teachers to send the assignment, student and rubric to support when the marking drifts. The honesty is to its credit, and the same variability applies to every AI marker on the market. Your job is to decide how much drift your assessment policy can absorb.

• Does it mark against your rubric and your syllabus outcomes, or does it make you rewrite the task to fit its templates?
• Who sees the mark first, and can you hold it back? Confirm the setting exists before you need it.
• Can you override a mark, and does the override leave a record of your reasoning?
• What happens to student work: where is it stored, is it used to train models, and how long is it kept?
• Does it connect to the platform your school actually runs, or will someone be retyping marks by hand?
• Does it read handwriting, if your assessments are still on paper?
• Is the feedback written in a register your students can act on without you translating it?

How does a teacher-first marking assistant compare?

A different design starts from the teacher’s pile rather than the student’s screen. The teacher brings in completed work, the AI drafts marking and comments against the department’s rubric and comment bank, and nothing reaches a student until a teacher has read and edited it. Jeddle is built that way, which suits departments whose bottleneck is the marking stack rather than the number of practice attempts in a lesson.

The trade-off runs both ways and is worth stating plainly. Student-facing instant feedback puts more revision cycles into a lesson than any teacher can mark by hand, and for formative practice that is the stronger design. Teacher-first drafting gets you consistency across thirty scripts and a mark you have personally signed off, which is what a moderation meeting actually requires. JeddAI drafts against your own rubric, success criteria and comment banks, and you edit every draft before release.

For Australian departments there is a second difference worth naming. Marking tools built here start from ACARA and state syllabus outcomes rather than AP task types, so a Year 11 English teacher is not translating an American rubric before the first script is marked. If expert essay marking on submitted work is the job you want lighter, that is the shape of tool to look at. If in-lesson practice is the job, the student-facing design wins, and there is no rule against running one of each.