Skip to main content
Back to home

Features

Controlled, reviewable AI assessment, end to end

Build the assessment, let the AI mark each answer against your rubric with the evidence cited, and hold every result until an assessor releases it. Built for RTOs, L&D and HR teams.

The essentials

Three features that do the heavy lifting

Core features

01

Marked against your rubric, evidence on every line

Every answer is scored against your own criteria, with the part of the answer behind each score quoted and the level explained. Nothing is released until your assessor signs it off. The AI reads their actual answers, not just their score.

AI generated

Marcus Hale

Programme Lead

75/100

Communication88%
Decision-making73%
Product knowledge64%
Strength - Marcus cites concrete examples on Q2 and Q6, showing he reflects on outcomes, not just intent.
02

Conditional branching

Show or hide questions based on previous answers. Build adaptive assessments that feel personalised to each respondent, reducing abandonment and improving data quality.

Every path is yours to draw

03

Custom branding

Add your logo, brand colours, and fonts. On Pro, remove Scorafy branding entirely so respondents see only your organisation on the assessment and the report. Your assessments, your brand.

Your logo.Your branding
BrandMidAccentPaper

Reports in your house style

Built for regulated cohorts

Controls you can point at

Scorafy is in active design-partner work with a not-for-profit serving 17+ countries across Education, Vocational Training, and HR. The controls below came out of that engagement and the assurance work that followed it. They are things you can show a privacy or AI reviewer - not a certification, and not a statement that Scorafy makes your organisation compliant.

01Delete media on submit
Opt-in per assessment. When it is on, candidate audio and video files are deleted from storage at submission, for every answer that already has a transcript. The transcript is kept for grading; the recording is not, and the deletion is recorded per answer and cannot be undone. An answer whose transcription did not succeed keeps its recording so the response can still be assessed.
02Reviewer score override
An assessor can override the score on any individual question, with a comment explaining why, and totals recompute automatically. The AI-generated score is kept alongside the override rather than replaced, so both numbers stay on the record as an audit trail.
03AI transparency notice
Tell candidates clearly that AI is used to analyse their responses, and publish a custom notice so anyone not comfortable can request an alternative method from their instructor. New results stay unreleased until an assessor releases them, so a candidate sees nothing until a person acts.
04Audio or video candidate choice
Owners can let candidates choose audio-only instead of video for any video response question. Lower biometric footprint, better accessibility, and identical transcription and grading either way.
05One-click audit pack
Export an assessment’s whole evidence chain as a single PDF and JSON pair - the model and configuration versions in force, the override ledger, and the release chain showing who signed off and when. It is the document you hand a moderator or an auditor when they ask how a mark was arrived at, and it needs no login to read.
06Confidence on every mark
Each question comes back with the model’s own certainty about that judgement, alongside a per-criterion explanation of which rubric level was awarded and why. Reviewers sort by confidence and work the borderline marks first, rather than reading every result at the same depth.
07Your own agreement rate
Every organisation gets an Accuracy page measuring how often its assessors agree with the AI - overall, per question, and per assessor. It surfaces which questions attract the most overrides and how each assessor moves scores, so you can fix a rubric that is producing disagreement rather than argue about it.
08Segregation of duties
The person who builds an assessment does not have to be the person who releases its results. Roles are separable across your team, and each release is recorded against the individual who made it.
09Feedback now, score after sign-off
An assessment can share its written feedback with a respondent immediately while holding the numeric score until an assessor confirms it. The score is withheld on the server - it does not reach the browser, and it is kept out of the AI’s written prose as well, so it cannot leak back through the wording.

Live and paced assessment

For when the whole cohort is in the room

Sending a link and waiting is one way to assess. The other is a room, a clock, and an assessor who decides what everybody can see. Scorafy runs both - and in a live sitting, the pacing you set is enforced by the server rather than trusted to the browser.

Live sittings are available on every plan, including free.

Assessor-paced live sittings

Run a whole room through an assessment together. You open one question at a time and every candidate answers it at the same moment - nobody races ahead, nobody sits on question nine while you brief question one. Switch any assessment into assessor-paced mode; the self-paced link keeps working exactly as before.

Sequencing enforced on the server

The content of an unopened question is never sent to a candidate’s browser, and the platform refuses an answer to a question that has not been opened. Read-ahead is closed at the boundary, not hidden with CSS - so the control holds even against a candidate reading the network tab.

Instruction and pause-and-discuss screens

Insert full screens between questions to brief a section, set a scenario, or hold the room for a discussion beat. They carry your own text, they do not count towards the question total a candidate sees, and they let a live sitting be taught rather than just administered.

Join by QR code or link

Put the QR code on the room screen and candidates are in within seconds - no accounts, no installs. The roster fills in live as people join, so you can see the room assemble before you open the first question.

Incremental server-side saving

Answers are written to the server as candidates type, not held in the browser until submit. A dropped connection, a flat battery, or a closed lid costs nobody their work.

Rehearse before the room arrives

Drive a complete live sitting against yourself first - open questions, answer as a candidate, watch the pacing - so the first time you run one is not in front of a cohort.

Forward-only mode

Stop respondents going back to revise an earlier answer once they have moved on. Useful for reaction-under-pressure questions and for any sitting where an answer is meant to be a first take.

Recording attempts, limited server-side

Cap how many times a candidate can re-record a video or audio answer. The limit is counted on the server, so it holds regardless of what the browser is told.

Self-paced assessment is unchanged - links, timers, scheduling and branching all work exactly as they did. Live sittings are an additional mode on the same assessment, not a different product. See how live sessions work

The full toolkit

Everything else you might need

Code answers

SQL, Python and other code answers are read and marked against your rubric, with the evidence quoted like any other answer. Code is read, not run.

CSV reference data

Attach a CSV dataset to a capstone task and the AI checks the learner's figures against it when it marks.

Answer signals

Assessors see when text was pasted, how long each question took and whether the candidate left the page. We don't run an AI-writing detector.

Private ideal answer and Q&A sheet

In a live session, read the ideal answer for each question on your own screen while candidates see only the question. Print the questions and ideal answers as a Q&A sheet.

Review requests

Switch it on and each released result ends with a request form. Respondents can ask for a review, and every request is tracked to a decision under Appeals.

Cohorts within one assessment

Give each cohort its own link, open date and close date, without duplicating the assessment.

Timed assessments

Set a time limit per assessment or per question. Respondents see a live countdown and unfinished answers auto-submit at zero. The whole-assessment limit is enforced server-side; per-question limits pace each question on screen.

Cohort reports

Aggregate a whole cohort into one report - score distribution, common strengths, shared gaps, and recommendations for the programme, built from the individual AI evaluations.

Response analytics

Track completion trends, score distributions, and dimension performance across all your assessments from one dashboard.

Assessment scheduling

Set open and close dates for your assessments. Respondents see a countdown before it opens and a closed message after.

Batch respondent import

Upload a CSV of names and emails to pre-populate your respondent list. Scorafy validates and deduplicates automatically.

PDF and CSV export

Export individual AI reports as branded PDFs, or download all response data as CSV for your own analysis.

Assessment duplication

Clone any assessment with one click - questions, options, scoring, and settings all carried over. Perfect for recurring programmes.

Webhook notifications

Point a URL at Scorafy and get HMAC-signed POSTs for four events - response completed, report released, assessment published, and assessment archived - with automatic retries. Report released fires only after an assessor releases the result.

API keys

Generate scoped API keys on the Pro plan and pull your assessments, responses, and released reports through the public REST API (v1). Keys are hashed at rest and revocable at any time.

Team roles and permissions

Invite team members with admin, member, or viewer roles. Control who can create assessments, manage billing, and view reports.

Enterprise-grade security

TLS encryption, row-level data isolation, HMAC-signed webhooks, and hashed API keys. Your data never trains AI models.

Dark mode

Switch between light and dark themes with one click. Follows your system preference by default.

Video and audio responses

Respondents record answers via webcam or microphone. Audio is transcribed automatically, and the AI analyses both the transcript and written responses together. Candidates can choose audio over video where the owner permits it.

File upload questions

Ask respondents to upload documents, resumes, or PDFs. Scorafy extracts the content and feeds it into the AI analysis alongside their other answers.

Weighted rubrics

Define specific criteria and performance levels. The AI assesses each respondent against your rubric with evidence from their answers - not generic feedback.

Grading schemas

Choose from built-in grading schemas (A-F, Pass/Fail, Competency) or create your own. Every report includes a grade mapped to your standards.

Resume cross-referencing

AI compares uploaded documents against assessment responses. Ideal for interview prep, competency validation, and portfolio review.

Canvas gradebook export

Export released results as a Canvas-compatible CSV that imports straight into the gradebook, plus a plain CSV for everything else. Marking in Scorafy does not mean re-keying grades into your LMS by hand.

Rubric import from documents

Upload the rubric you already have - a Word document, a PDF, a spreadsheet - and Scorafy reads the criteria and levels out of it into a working rubric. You correct it rather than retyping it.

Terminology-aware marking

Tell an assessment that regional or professional wording variants mean the same thing - UK, US, Israeli and Latin American terms for the same procedure, say - and answers are marked on meaning rather than exact vocabulary. Candidates are not penalised for the words their training used.

One-click template activation

Pick a ready-made template and it builds the real assessment - questions, rubric, scoring, settings - into your account, ready to edit. Not a preview you then have to recreate by hand.

Audio transcription

Automatic speech-to-text for audio and video responses powered by Deepgram Nova-3. Transcripts are included in the AI analysis for richer, more accurate reports.

Plans

Feature comparison

See exactly what is included in each plan. One AI evaluation is one completed respondent assessment, evaluated.

FeatureFreeStarterGrowthProEnterprise
Active assessments1315UnlimitedUnlimited
AI evaluations per month101005002,000Custom
Questions per assessmentNo limitNo limitNo limitNo limitNo limit
AI evaluation reports
All 26 question types
Timed assessments (server-enforced)
Assessor-paced live sittings
Instruction and pause-and-discuss screens
Join by QR code or link
Confidence and per-criterion explanations
Release gate and reviewer overrides
Feedback now, score after sign-off
One-click audit pack (PDF and JSON)
AI-vs-assessor Accuracy page
Conditional branching
Assessment scheduling
Assessment duplication
CSV export
Weighted scoring
Custom branding--
PDF report export--
Batch respondent import
Response analytics
Cohort reports
Team members1310UnlimitedCustom
Unbranded assessments and reports---
Public API (v1)---
Webhooks---
Video and audio responses
File upload questions
Weighted rubrics
Grading schemas
Resume cross-referencing
Audio transcription
Data residency options----
Configurable data retention
Dedicated support----
Invoice billing----

Enterprise

Running assessments across a whole organisation?

Enterprise gives you everything in Pro plus data residency options, configurable retention, dedicated support, and invoice billing. Evaluation volume and seats are scoped to your organisation - talk to us for a quote.

See it live

See these features in action

Try the interactive demo - no sign-up required.

Free plan included · No credit card required