Testam is a complete assessment platform — a reusable question bank across ten question types, exams with the controls real assessments need, proctoring you configure, grading that finishes itself, and live Kahoot-style quizzes drawn from the same material. Free to use, with nothing held back behind a tier.
The value isn't in delivering a test — it's in the bank of reviewed questions you keep afterwards, and in knowing which of them actually measure anything.
Every type accepts markdown and media — in the question and in each individual option — so diagrams, snippets, audio, and video are first-class rather than bolted on.
One correct option, with per-option feedback.
Auto-gradedSeveral correct options, with partial credit.
Auto-gradedThe fastest thing to write and the fastest to answer.
Auto-gradedExact match, keyword list, or a regular expression.
Auto-gradedAccept a value, or anything inside a tolerance range.
Auto-gradedPair items across two columns, scored pair by pair.
Auto-gradedPut steps in sequence — position-aware scoring.
Auto-gradedKeyword or AI validation, with a human able to override.
AI or manualRubric-based grading, manual or AI-assisted.
AI or manualSyntax-highlighted input, graded against your criteria.
AI or manualFive steps, and only the first one is writing. Everything after it is configuration you set once and reuse.
Write questions once — or generate a draft set with AI — and file them into a category tree. Markdown and media work in the question and in every option.
Questions move draft → pending review → approved. Moderators see exactly what's waiting on them, so nothing half-finished reaches a candidate.
Pick questions or whole sections, set marks and per-question time, choose navigation, shuffling, attempt limits, and how access is granted.
Candidates take it on any device with auto-save and a server-side timer. Proctoring signals, if you enabled them, are recorded as they happen.
Objective answers are already scored. Subjective ones go to a grader or to AI. Then release scores, answers, and certificates on your schedule.
Most of a paper grades itself. The part that genuinely needs judgement is the only part that reaches a person — and AI can draft even that.
Scored the moment the attempt is submitted — partial credit, negative marking, tolerance ranges, and position-aware ordering all included.
AI reads your rubric and proposes a score with its reasoning shown. It is a first pass, not a verdict — a grader can override any of it.
A grading queue with the rubric beside the answer, inline comments, and the option to grade one question across every candidate at a time.
Scores, correct answers, and explanations are released on your schedule — immediately, after the deadline, after manual grading is finished, or never.
Every measure is a per-exam switch, and candidates are shown which ones are active before they begin. Signals are written to a timestamped violation trail you review beside the attempt.
We won't pretend browser proctoring makes cheating impossible. It raises the cost and gives you evidence to look at — treat the log as an input to a human decision, not as a verdict.
Require fullscreen to start. On exit: warn, count it, auto-submit after N, or lock the attempt.
Every time the window loses focus is timestamped, with a threshold you set.
Block the shortcuts outright, and log each blocked attempt.
Detect when devtools are opened during an attempt.
Allow or refuse desktop, tablet, and mobile independently per exam.
Off unless you turn it on. Candidates are told before the attempt begins.
Turn any set of approved questions into a real-time room. It runs over Firestore, so the leaderboard moves the moment an answer lands rather than when someone refreshes.
The useful signal is which questions separated the people who knew the material from the people who didn't — and Testam measures that for you on every sitting.
An organization has four roles, and a question can't reach a candidate until someone other than its author says it's ready.
Owns the organization — members, roles, branding, settings, and everything below.
Reviews and approves submitted questions, and grades attempts that need a human.
Writes questions and builds exams. Submits them for review rather than publishing directly.
Takes exams and sees their own results and certificates — and nothing that isn't theirs.
No tiers, no add-ons, no per-candidate charge. If it's here, it's yours.
Ten question types, markdown and media in both the question and every option, hierarchical categories, variants, cloning, and a draft → review → approved workflow so nothing half-written reaches a candidate.
Practice, assessment, certification, survey, or live quiz. Schedule windows, cap attempts, shuffle questions and options, and choose sequential, free, or section-based navigation.
Open links, passcodes, an email whitelist, or a one-time passcode over email or SMS. Invite manually, upload a CSV, or open registration to anyone with the link.
Fullscreen enforcement, focus and tab-switch tracking, copy-paste, right-click and devtools blocking, and optional webcam or screen recording — every signal written to an auditable violation trail.
Objective answers are scored on submission with partial credit and negative marking. Subjective answers go to a grader, to AI, or to AI-then-human — with the rubric and the reasoning shown.
Score distributions and timing per exam, difficulty and discrimination index per question, strengths and weaknesses per candidate, and roll-ups across the whole organization.
Turn any set of questions into a Kahoot-style room. Players join with a code, answers stream in over Firestore, and the leaderboard moves as they land.
Generate a branded PDF automatically on a pass, with a unique code and QR anyone can check. Set expiry, revoke when needed, and customise the template.
Exams, attempts, and results have to live on our servers for us to deliver and grade them. That's the trade — so here is exactly what we do with them.
Read the full privacy policy and the terms of use.
“We were running certification rounds out of spreadsheets and a video call. Testam replaced the whole thing in a week. The part that actually changed our process was the discrimination index — we retired eleven questions that everyone was getting right for the wrong reason.”
“Section-based navigation with per-section timers is the thing I couldn't find anywhere else for free. Code questions with markdown in the options mean I can finally set the questions I actually want to set.”
“Sending a screening assessment now takes about two minutes — pick the bank, set an OTP link, upload the candidate CSV. AI grades the written answers first and I confirm. What used to be an evening is now a coffee.”
“I use the same question bank two ways: a serious assessment on Friday, and a live quiz on Monday to see what stuck over the weekend. Not having to maintain two sets of material is the whole win.”
Yes. Testam is free to use — for individuals, teams, schools, and companies alike. There is no trial period, no per-candidate charge, and no feature locked behind a paid tier. You can run unlimited exams, questions, and attempts.
No. Testam is free to use but its source code is proprietary and is not published. You use the hosted application; you don't self-host or modify it.
Not necessarily. It depends on how you configure access. An exam can be opened to anyone with the link, gated behind a passcode, restricted to a whitelist of email addresses, or protected with a one-time passcode delivered by email or SMS. Only exams that need to be tied to a member of your organization require the candidate to sign in.
Ten: single-choice, multiple-choice, true/false, fill in the blank, short answer, long answer, matching, ordering, numerical, and code. Questions and every individual option support markdown and attached media, so diagrams, snippets, audio, and video all work.
Objective types are graded the instant an attempt is submitted, including partial credit and negative marking where you've configured them. Short answer, long answer, and code go to a grading queue where you can grade manually, ask AI to grade against your rubric, or use AI as a first pass that a human then confirms or overrides. AI never has the last word unless you decide it does.
Proctoring is off by default and configured per exam. When enabled it can require fullscreen, detect and count exits, track window and tab focus loss, and block copy-paste, right-click, and developer tools. Webcam snapshots and screen recording are separate options you must turn on deliberately. Every signal is recorded as a timestamped violation you can review alongside the attempt — and candidates always see which measures are active before they begin.
No tool can, and we won't claim otherwise. Browser-based proctoring raises the cost of cheating and gives you an evidence trail to review; it is not a guarantee. For high-stakes certification, treat the violation log as one input into a human decision, not as a verdict.
A live quiz is host-driven and synchronous. You open a room, players join with a short code from any device, and you advance the questions yourself while answers stream in and the leaderboard reorders in real time. It draws from the same question bank as your exams, so material you built for an assessment can be replayed as a warm-up or a revision session.
Yes. An organization has roles — admin, moderator, instructor, and member. Instructors create questions and exams, moderators review and approve them, and admins manage everything including users and settings. Questions move through draft, pending review, and approved states so a reviewer sees exactly what is waiting on them.
Exams, questions, attempts, and results are stored on our servers so they can be delivered and graded — this is a hosted platform, not a local tool. Within your organization, access is enforced by role: members see only their own attempts. We never sell your data, never use it for advertising, and never use it to train models.
Only when you invoke them, and only what the request needs. AI question generation sends the topic or prompt you supply. AI grading sends the question, the rubric, and the candidate's answer. Nothing is sent automatically, and account credentials are never included. If you don't use these features, no content leaves the platform.
Yes. Organizations can upload a logo and set branding that appears across exam pages and generated certificates. Certificate templates are customisable, and each issued certificate carries a unique verification code and QR that anyone can check without an account.
Yes — the whole platform is responsive, and candidates can take exams on a phone. You can also restrict an exam to specific device classes if, for example, a coding assessment only makes sense on a desktop.
Yes. Questions, exams, and results can be exported, and you can delete any content or your entire account at any time. Nothing you create is held hostage.
Answers auto-save as the candidate works, and the attempt keeps its own server-side timer. Reconnecting resumes the attempt where it left off, with the remaining time intact — a dropped connection doesn't hand back extra minutes, and it doesn't lose work.
Write a handful of questions, set how people get in, and send the link. Everything — the bank, the proctoring, the grading, the analytics, the certificates — is free, with nothing waiting behind an upgrade.
