Create rubric-graded activities, collect multi-modal submissions, and publish AI-powered evaluations calibrated with your own reference examples — all in one workflow.
From activity creation to published student results in four straightforward steps.
Set a title, write the prompt, choose the modality (writing, speaking, or file upload), select a rubric, and pick the AI coaching mode. Assign to individual students or groups.
Students write in the rich text editor, record audio directly in the browser, or upload a file. Submissions are saved automatically as drafts and submitted when ready.
The AI scores every criterion, generates per-criterion feedback, and calibrates against the most semantically similar reference examples from your library.
The teacher reviews the AI draft, edits any score or comment, then publishes. The student sees the full criterion-by-criterion breakdown with feedback immediately.
Shared rubrics across all activities in your organisation. Three criterion types give you full control over how each dimension is scored:
Global rubrics (platform admins) are available to all organisations. Per-criterion AI guidelines let you add scoring intent beyond the descriptor text.
Before evaluating any submission the AI searches your reference library using semantic vector similarity — not keyword matching. The most relevant examples (excellent, good, average, poor student answers or marking guides) are injected into the evaluation prompt as calibration anchors, keeping scores consistent across sessions, evaluators, and time. Add as many examples as you like; the AI always retrieves the most relevant ones for each submission.
Four distinct modes let you control the depth and focus of AI feedback:
Students can write in a full rich-text editor, record audio directly in the browser (auto-transcribed by Whisper for evaluation), or upload a file. Each modality feeds into the same rubric evaluation pipeline so the same rubric and scoring process applies regardless of how the student submitted their work.
Track evaluation outcomes across your organisation. See criterion-level averages (with percentage scores for easy comparison), identify your weakest criteria, monitor submission and evaluation rates per activity, and spot patterns across cohorts. Export everything to CSV for offline analysis or institutional reporting.
Write activity prompts and receive AI evaluations in English, Arabic, Spanish, French, Hindi, Chinese, or German. Speaking submissions are transcribed by Whisper with automatic language detection. Each language configuration applies independently to the prompt, the student-facing interface, and the AI evaluation output.
"The RAG reference examples changed everything. I uploaded three strong essays and two weak ones. Now the AI scores consistently match what I would give — it's not just generic AI feedback anymore, it's calibrated to my standards."
"Being able to record audio submissions directly in the browser and have them transcribed automatically meant I could finally run speaking assignments at scale. The AI gives per-criterion feedback even for spoken answers."
"The draft-then-publish flow is exactly right. The AI draft saves me 80% of the marking time, I review and adjust anything that needs a human eye, then publish. Students get detailed feedback within an hour of submitting."
Essay, report, and academic writing courses where rubric consistency and formative feedback at scale are essential.
IELTS, TOEFL, OET, and general EFL/ESL programmes needing structured writing and speaking assignments with fast turnaround.
Universities and colleges running assignment-based assessment where faculty need to maintain marking standards across large cohorts.
Private tutoring businesses that want to offer structured, evaluated homework and practice assignments to their students.
Give students fast, consistent, rubric-calibrated feedback at any scale.
See All Features View Pricing