Skip to main content
The full feature tour

Every capability, in one place. No asterisks.

Authoring, delivery, grading, certification, reporting, governance — the features Setec Millenium ships, organised the way your team actually thinks about training programmes.

Authoring

Author quizzes the way your team thinks.

Start from a PDF, a knowledge-base URL, plain text, or an existing question bank. Refine every draft before it ships.

Source-grounded AI generation

Drop in a PDF or a URL. The pipeline extracts structure, surfaces candidate questions, and lets you accept, edit, or reject each one. Nothing publishes without your review.

Five question types

Single-choice, multi-select, true/false, short text, long text. Each comes with hints, rationale fields, and reference-answer capture for the AI rater.

Existing bank import

Pull in CSV or JSON exports from your previous LMS. Difficulty ratings and answer keys travel with the question.

Reference answers & rationale

Every question can carry an authoritative answer, an explanation, and the source page or section it came from — surfaced to reviewers and used by the AI rater.

Delivery

Adaptive or linear, with the controls compliance wants.

Lock down time limits, locale, access mode, and the consent gate before a single participant sees the quiz.

Adaptive engine

Each question carries a difficulty rating. A correct answer escalates to a harder follow-up; an incorrect one eases back. Works for any question type except long-text.

Linear delivery

Pin questions in a fixed order with optional back-navigation. Suits certification paths where order is part of the contract.

Six interface locales

English, Dutch, German, French, Spanish, Russian out of the box. Participants can switch at any time; their progress stays intact.

Account or invite-link access

Restrict a quiz to signed-in users, open it through a tokenised invite link, or both. Tokenised links capture only the fields you ask for.

Grading & certification

Auto-score, calibrate, certify — and audit every step.

Objective answers grade in milliseconds. Written answers go to manual review or a calibrated AI rater. Passing attempts issue certificates with unique IDs.

Objective auto-scoring

Single-choice, multi-select and true/false are scored against the configured answer key, in milliseconds, on submission.

Calibrated AI rater

Written answers route to OpenRouter with the question prompt, the reference answer, and the rubric your team writes. You set the strictness and the confidence threshold; below it, the response escalates to manual review.

Manual review queue

Reviewers see the participant's response, the reference answer, and the rubric side by side. Approvals and change requests are timestamped and attributed.

Certificate PDFs

Passing attempts issue a certificate with a unique ID, the score, the issued date, and the issuer. Downloadable, printable, and signed for tamper evidence.

Reporting & analytics

Reports that survive a compliance review.

Completion rates, pass rates, score distributions, per-question drop-off, and per-cohort comparisons — exportable, filterable, audit-ready.

Completion & pass-rate dashboards

Track how many participants finished, how many passed, and how those numbers trend over time. Filter by cohort, locale, or quiz.

Score distribution

See how your cohort actually scored — the median, the tail, the bimodality. Spot quizzes that discriminate well from quizzes that don't.

Per-question drop-off

Find the questions that lose participants and fix them. We surface drop-off rates, time-on-question, and the most common wrong answers.

CSV export

Every report exports to CSV with stable column names, ready for the systems your organisation already runs.

Governance

Roles, audit, and the controls a security review wants.

Granular roles, immutable audit log, workspace isolation, and the paperwork to back it all up.

Four built-in roles

Org admin, quiz creator, org member, and participant. Custom roles on the Enterprise plan. Permissions scoped per workspace, not per request.

Audit log

Every action — quiz publish, attempt submission, review decision, certificate issuance — is logged with actor, timestamp, and payload diff.

Workspace isolation

Each tenant's data is scoped at the database level. RLS policies make cross-tenant reads impossible from a normal client. Admin clients are server-only by design.

DPA & procurement paperwork

Standard DPA, redlines welcome, security questionnaires on request. Enterprise tier includes a dedicated success engineer for the procurement conversation.

Platform

OpenRouter, OpenAPI, and a co-pilot in the corner.

Route structured AI work through OpenRouter so your admin can pick the model family and watch token spend in one place. Power users get an in-app co-pilot.

OpenRouter-backed AI

Structured completions route through OpenRouter, so you can pick the model family per workspace, watch token spend in real time, and switch providers without touching code.

Per-workspace controls

Cap AI spend per workspace, pin a model for compliance reasons, or set a nightly budget. Configurable from the admin dashboard — no engineering required.

In-app co-pilot

Role-aware agent runs alongside the workspace: read tools auto-execute, mutation tools wait for human approval. History persists across sessions.

Bring-your-own keys

Enterprise customers can route through their own OpenRouter account, with usage billed directly to them. Audit-friendly and cost-predictable.

See it run

Read enough. Try it on your own source material.

Sign up, drop a PDF into the source materials library, and watch the first draft of an adaptive quiz appear in your workspace.

Features — Setec Millenium — Setec Millenium