For getting a whole class ready for a test, seven tools cover the job between them, and no single one does all of it. Wayground runs the weekly retrieval routine and shows you who still does not know what; Conker builds the standards-aligned quiz when you have nothing to start from; Brisk Teaching turns the material you already have into review questions without retyping it; MagicSchool writes the study guide at more than one reading level; Class Companion is the one that practices extended written answers, which quizzes cannot touch; Khanmigo picks up the students who are furthest behind; and Knowt is what you hand students for their own time.
The part worth knowing before you choose: the two review methods with the strongest research behind them, quizzing and spacing, are the two these tools automate most cheaply, and four of the seven do the core work on a free teacher plan.
British readers searching for revision tools are in the right place. This page uses the US terms, so "test prep" and "review" here mean what "revision" means in the UK.
The five stages of a review cycle, and which tool fits each
Most tool lists sort by feature. Review does not work that way. It works as a sequence, and a tool that is excellent at stage three is useless at stage one. Sorting the field by stage is what makes the choice obvious.
Stage 1, diagnose. Find out what the class actually does not know, as opposed to what you assume they forgot. A short low-stakes quiz with per-question data. Wayground and Conker both do this.
Stage 2, build. Produce the review materials: question banks, study guides, practice sets, at the levels your class actually reads at. Conker, Brisk Teaching and MagicSchool.
Stage 3, drill. Re-expose students to the content repeatedly over days rather than once the night before. Wayground, Knowt.
Stage 4, write. Practice the extended responses that carry most of the points on AP, state and end-of-course tests, with feedback fast enough that students revise while they still remember what they meant. Class Companion.
Stage 5, catch up. Handle the students who are two years behind and need something other than more of the same review. Khanmigo.
A teacher who owns one tool per stage has the cycle covered. A teacher who owns three tools that all do stage two has a problem this article is meant to prevent.
The comparison at a glance
| Tool | Best use | Strengths | Limitations | Price / free tier (checked Sept 2026) | Privacy considerations |
|---|---|---|---|---|---|
| Wayground | Running retrieval practice as a class routine | Live and student-paced delivery, per-student mastery data, 25+ accommodations including Read Aloud | Free plan caps at 100 students and 20 saved resources, limited question types | Basic free for every teacher, forever; Individual (Super) price not published; school and district by quote (plans) | Directory record blank on FERPA, COPPA, hosting and minimum age; students can join by code without individual accounts |
| Conker | Building a standards-aligned quiz from nothing | Multiple question types, ready-made NGSS and TEKS banks, exports to Google Forms and Canvas | Paid pricing not readable on its own pricing page; no directory rating yet | Free to try; paid tiers unverified (pricing) | Directory record blank on all four fields; has a 9ine assessment on record; Google Forms export means students need never enter Conker |
| Brisk Teaching | Turning existing material into review questions | Works inside Docs, Slides, YouTube and your LMS; quiz generator on every plan; 50+ languages | Browser extension only; batch feedback has upload limits on free | Teacher plan $0 with unlimited usage; Premium and Intelligence custom (plans) | Directory record blank on all four fields; it reads whatever document is open, so the district question is what leaves the machine |
| MagicSchool | Study guides at more than one reading level | 80+ teacher tools, quizzes and class writing feedback on free | Free plan has generation limits; student tools and learning insights are paid | Free forever; Plus $8.33/user/month billed annually or $12.99 monthly; Enterprise custom (pricing) | Directory record blank on all four fields; custom data privacy agreements are an Enterprise feature |
| Class Companion | Practicing written exam responses | Instant feedback on extended writing; names CAASPP, STAAR and AP prep explicitly | Only one directory review so far; school pricing by demo only | Teacher tier free; school and district by demo (plans) | Vendor states it complies with FERPA, COPPA and state privacy laws and does not sell or rent personal information; directory record is blank and currently wrong (see Safety) |
| Khanmigo | Students furthest behind the class | Standards-tied practice, progress summaries, district SSO via Clever SecureSync or ClassLink | Student use needs accounts; free tier is teacher-side only | Free for teachers; $4/month or $44/year for learners and families; districts custom (pricing) | Directory record blank on all four fields; governed by Khan Academy's privacy policy |
| Knowt | What students do on their own time | Unlimited practice tests, spaced repetition, teacher classes and student analytics on free | Rostering and LMS integration are paid; on free, students self-register | Free for teachers, capped at 20 "ultra features" a month; Ultra price not published (teachers) | Directory records FERPA yes, COPPA yes, US hosting, no minimum age, which is more than most of this field states |
Wayground: best for running retrieval practice as a weekly class routine
Wayground is the best fit for making review a routine rather than an event, because it is the only tool here that delivers the same question set live or student-paced and returns per-student mastery data from both. That is what turns a review quiz from a grade into a diagnostic. It carries a 4.71 rating from 7 reviews on the AI Educator Tools directory.
The platform was renamed from Quizizz, and its own help center still signs off as "Quizizz Inc. (DBA Wayground)", so teachers searching under either name are looking at the same product. The free Basic plan includes assessments, lessons, passages, flashcards and videos, live and student-paced delivery, AI generation and import, 25+ accommodations including Read Aloud, and live instruction data through Classroom Pulse (plans).
Watch out for: the free plan's ceilings, which are the most consequential limit on this page and the least advertised. Basic caps at 100 students and 20 saved resources, with limited question types and limited access to the standards-aligned library (plans). A high school teacher with five sections of 28 passes the student cap before the first review session, and 20 saved resources is roughly one unit. Elementary teachers with a single class will not notice either limit.
Pricing: Basic is "Free for every teacher, forever. No credit card, no commitment." The Individual (Super) plan adds premium question types, unlimited storage, team sharing and student growth tracking; its price is not published on the help center article that describes it, which points to a separate pricing page. School and district plans are by quote. Checked September 23, 2026 (plans, Individual plan).
Safety: the directory record states nothing about FERPA, COPPA, data hosting location or minimum age, which is a gap worth raising with your district before a rollout. The structural point in Wayground's favor is that students can join a session with a code rather than an individual account, which is the lower-data route through a review lesson. Whether your district needs a signed data agreement before any student answers are stored is a question for your district, not for the vendor's marketing.
Conker: best for building a standards-aligned quiz from nothing
Conker is the best fit for the moment you need a quiz on a standard you have no materials for, because it ships ready-made K-12 banks aligned to NGSS and TEKS and generates across multiple question types from a topic and a difficulty setting (Conker). Where Wayground is the delivery routine, Conker is the supply.
It exports to Google Forms and Canvas, which matters more than it sounds: it means the quiz can run entirely inside tools your district has already approved, with students never creating a Conker account.
Watch out for: the absence of a directory rating. Conker has no reviews on the directory yet, so unlike Brisk Teaching at 35 reviews or MagicSchool at 19, there is no weight of educator experience behind it here. Treat it as promising rather than proven.
Pricing: the site offers a free trial, and the directory records the product as freemium. Conker's own pricing page did not render its plan table when it was fetched on September 23, 2026, so the paid tiers are unverified and you should read them yourself before committing a department budget (pricing).
Safety: the directory record is blank on FERPA, COPPA, hosting location and minimum age, though Conker does have a 9ine assessment on record, which more than half this field does not. Conker's own privacy policy, dated February 24, 2023, states that the company complies with "FERPA, COPPA, GDPR and the Australian Privacy Act", that it does not "sell, rent or lease your information", and that for users under 18 it strips email addresses and device IP addresses from third-party data sharing and may shorten retention. Those are the vendor's statements, and your district still confirms them in its own agreement. The practical mitigation is the export path: run the quiz in Google Forms and the student data stays where your district already governs it.
Brisk Teaching: best for turning material you already have into review questions
Brisk Teaching is the best fit for the teacher who does not need new content, only a way to convert eight weeks of existing material into review items, because it works as a browser extension on top of Google Docs, Slides, YouTube and your LMS rather than asking you to paste anything into a new app. It is the most reviewed tool on this page, at 4.77 from 35 reviews on the directory.
The quiz generator is on every plan, including free, which is unusual in this field. So is the feedback set: glow and grow, next steps, rubric criteria feedback and feedback insights all appear on the free teacher plan (plans).
Watch out for: the extension model itself. Brisk works where you already are, which is its strength and its constraint. It is a Chrome and Edge extension, so a district that locks down extension installs has to approve it centrally before a single teacher can use it, and there is no browser-free route in.
Pricing: the teacher plan is $0 with unlimited usage, 20+ creation tools, generation in 50+ languages and the Brisk Next content hub. Batch feedback carries upload limits on free. Premium (schools and districts) and Intelligence (districts) are custom priced. Checked September 23, 2026 (plans).
Safety: the directory record states nothing about FERPA, COPPA, hosting or minimum age, but Brisk's own privacy notice, effective September 7, 2026, answers the question a district technology director will actually ask. It states that "Student Data is owned and controlled by the School" and that Brisk acts as a processor, and that it does "not use or disclose Student Data for targeted advertising purposes". It is also specific about what the extension sends: where the AI Enhancement feature is turned on for supported third-party AI platforms, Brisk collects the prompts submitted, the responses returned, which platform was used, the page URL and session duration and usage metrics. That is more disclosure than most of this field offers, and it is also a bigger collection surface, so read it before a department-wide rollout rather than after.
MagicSchool: best for study guides at more than one reading level
MagicSchool is the best fit for producing the review handout a mixed-attainment class can actually read, because leveled rewriting is one of the 80+ tools on its free plan rather than a paid feature. A study guide that half the class cannot decode is not review, and this is the cheapest fix for that on the page. It holds 4.74 from 19 reviews on the directory.
Quizzes and class writing feedback are also on the free plan (pricing), which makes MagicSchool a reasonable single starting point for a teacher who wants to try one tool rather than five.
Watch out for: generation limits on the free plan, which MagicSchool confirms exist without publishing the number. If review season is the time you generate most heavily, that is exactly when you will meet the ceiling. Student-facing tools and student learning insights are also paid, so the free plan is a teacher's workshop rather than something students log into.
Pricing: Free is $0 forever. Plus is $8.33 per user per month billed annually, or $12.99 billed monthly, and adds unlimited generations, output history, Studio Mode editing, 50+ student tools and student learning insights. Enterprise is custom. Checked September 23, 2026 (pricing).
Safety: the directory record is blank on FERPA, COPPA, hosting and minimum age. MagicSchool's free plan advertises "best in class privacy and security", which is the vendor's phrase and not an independent finding. The concrete detail that matters to a district is further down the same page: custom data privacy agreements are an Enterprise feature. If your district requires a signed DPA before any tool touches student work, the free plan will not satisfy it and the budget conversation starts earlier than you expected.
Class Companion: best for practicing written exam responses
Class Companion is the best fit for the part of test prep that quizzes cannot reach, because it gives students instant feedback on extended written answers and names CAASPP, STAAR and AP preparation directly (Class Companion). On tests where the constructed response carries most of the points, this is the stage most review cycles skip, and it is skipped because hand-grading thirty practice essays twice is not possible in a two-week window.
The mechanism that makes it work is speed rather than accuracy. Students receive feedback immediately after submitting, which means they revise while they still remember what they were trying to say. A comment returned nine days later is read as a grade, not as instruction.
Watch out for: the thin evidence base here compared with the rest of the page. Class Companion has one review on the directory, against 35 for Brisk Teaching. Also, treat AI feedback on writing as a first pass that students respond to, not as a score. If you are looking for a wider field on this specific job, the AI writing feedback tools compared for teachers covers it in more depth.
Pricing: the Teacher tier is free and includes instant AI feedback, basic insights, the public content library and in-app chat. School and district plans add LMS integration, AI and plagiarism detection, a private content library and dedicated support, and are priced only after a demo, so there is no public figure. Checked September 23, 2026 (plans).
Safety: Class Companion states on its own site that it "is compliant with FERPA, COPPA, and all state privacy laws" and that it does not "market, sell, or rent any user personal information, or use the data for advertising" (Class Companion). Those are the vendor's statements, and a district still has to confirm what its own agreement says. One caution specific to our directory: the Class Companion record on aieducator.tools currently carries another vendor's plan names and privacy policy link, so read the vendor's own pages rather than the directory record for this tool until that is corrected.
Khanmigo: best for the students furthest behind the class
Khanmigo is the best fit for the students a class review will not reach, because it ties practice to standards and returns progress summaries, which is what you need when a student is working two grade levels below the test rather than three lessons behind it. It holds 4.40 from 5 reviews on the directory.
Reviewing a unit with a class assumes the class has the prior knowledge the unit rests on. For a handful of students in most rooms, that assumption fails, and more review of the current unit does not fix it. Khanmigo's value in a test-prep cycle is that it can run underneath, on the prerequisite content, while the rest of the class works on the unit.
Watch out for: the split between what is free and what is not. Khanmigo is free for teachers, and the teacher-side features are lesson planning, progress summaries, rubrics and exit tickets. Student tutoring is the paid side, at $4 a month for families or a district license. So the free plan does not put Khanmigo in front of the students this section is about. That gap is the thing to check before you build a plan around it.
Pricing: free for teachers. Learners and parents pay $4 per month or $44 per year. Districts are custom priced and get Khanmigo for all teachers and students, district and school data, SSO through Clever SecureSync or ClassLink, and a dedicated customer success manager. Checked September 23, 2026 (pricing).
Safety: the directory record states nothing on FERPA, COPPA, hosting or minimum age, and the Khanmigo pricing page carries no minimum age either. Khan Academy's privacy policy is the governing document. The district route, with Clever or ClassLink SSO, is also the route where your district's own agreement applies, which is generally what a technology director will prefer over teachers signing students up individually.
Knowt: best for what students do on their own time
Knowt is the best fit for the handoff at the end of a review cycle, because its free plan gives students unlimited practice tests and spaced repetition, which are precisely the two methods with the strongest research support, at no cost and with no purchase decision. It carries a 5.0 rating from a single review on the directory, and its record is the most complete on this page: FERPA yes, COPPA yes, hosting in the United States, no minimum age.
Teachers can create or join classes and view student analytics on files on the free plan, capped at 20 "ultra features" a month (teachers). Knowt's own line is worth quoting: "We don't want educators to pay, but schools or districts can purchase licenses."
Watch out for: rostering. Clever, ClassLink and SIS integration, along with LMS integration and incident monitoring for violence, safety and profanity, are on the paid Ultra plan. On the free plan, students self-register, which means a teacher is asking a class to create accounts on a service the district has not provisioned. In a district with an approved-tools list, that is the step that stops this.
Pricing: free for teachers, including classroom AI tools, flashcards with study modes, autograded assessments and class analytics, limited to 20 ultra features a month. The Ultra plan adds unlimited AI usage on teacher-owned files, rostering, LMS integration and extra question types; its price is not published on the teachers page. Checked September 23, 2026 (teachers).
Safety: the directory record states FERPA yes, COPPA yes, United States hosting and no minimum age. "No minimum age" is a statement about the product's terms, not a judgment that it suits any age, and under COPPA a service used by under-13s in school normally runs on the school-consent route described below. For students working independently, the AI study tools a teacher can hand a high school student goes further into what is safe to recommend for personal use.
What about Quizlet, ChatGPT, Gemini Notebook and Diffit?
These four come up in almost every search on this topic, so here is where each lands for class review, in one line.
Quizlet is still the default students already use, and for flashcard drill that familiarity is worth something; for teacher-side class data it gives you less than Wayground or Knowt.
ChatGPT will write you a practice test in thirty seconds and is genuinely good at it, but it holds no class data, aligns to no standard unless you supply it, and is not a tool you point students at without a district position on it first.
Gemini Notebook, the tool Google renamed from NotebookLM in July 2026, is the strongest option on this list for turning your own unit materials into a study guide grounded in those materials rather than in the model's general knowledge, which is the failure mode of generic AI review content.
Diffit is the better choice than MagicSchool when the specific job is leveling one reading passage several ways, and the weaker choice when you want the study guide, the quiz and the feedback in one place.
If the narrow job is the quiz artifact rather than the review cycle, the AI quiz makers compared for teachers is the more direct page, and the AI flashcard generators compared on price, data hosting and under-13 use covers the drill artifact.
A two-week review plan for a tenth-grade biology unit test
This is a worked plan for a unit test on cell division and genetics, in a 10th-grade class of 28 with a 50-minute period, using four of the seven tools. It is written as a plan to follow, not as a report of a class that ran.
Day 1 (Monday, week one), diagnose. Run a 12-question Wayground set covering the whole unit, student-paced, 15 minutes, ungraded and announced as ungraded. The output you want is the per-question breakdown, not the scores. Expect two or three questions where the class is below 50 percent. Those are the unit, for the next two weeks. Everything else gets one pass.
Day 2, build. Take the two or three weak standards into Conker and generate a question bank on each, more items than you will use. Separately, run MagicSchool's leveling on the unit summary to produce the study guide at two reading levels, and give both versions to the whole class rather than assigning them, so no student is publicly sorted.
Day 3, first drill. Ten minutes at the start of the period, live Wayground, drawing only on the weak standards. Then teach the rest of the period as normal. The ten minutes is the intervention; do not extend it.
Day 5 (Friday), write. One Class Companion extended response on the weakest standard, done in class under time. Students get feedback the same period and revise before they leave. This is the stage most plans skip, and skipping it is why students who can answer multiple choice on a topic still lose the constructed response points.
Day 6 (Monday, week two), space it. Second Wayground drill on the same weak standards, deliberately three days after the first rather than the next day. The gap is doing the work here, and closing it to reduce the wait would remove the effect the IES practice guide is describing.
Day 8, write again. Second Class Companion response, this time on a standard the class was middling on rather than worst at. Students who revised on day 5 should visibly transfer it.
Day 9, catch up. For the three or four students whose day 1 data showed a prerequisite gap rather than a unit gap, assign Khanmigo practice on the prerequisite content while the rest of the class does a third Wayground drill. This is the only differentiated day in the plan and it is differentiated by diagnosis, not by group.
Day 10 (Friday), hand off. Give students the Knowt set built from the unit materials, and say plainly what it is for: spaced practice over the weekend and the days before the test, not a cram the night before. Then stop assigning review. The last useful thing a teacher does in a review cycle is get out of the way.
Total teacher time after setup: roughly 10 minutes of class time on four days, one period of writing practice twice, and no hand-marking of practice essays.
Three questions that pick the tool for you
1. Do your students have accounts on a district-approved platform, or not? If they do not, the tools that work by join code or export (Wayground, Conker through Google Forms) are the only ones you can actually use this month. Everything else needs provisioning, and provisioning takes longer than your review window.
2. Is the test mostly selected response, or mostly extended writing? Selected response points to Wayground, Conker and Knowt. Extended writing points to Class Companion, and the tools that drill recall will not move the score much on its own.
3. Is the problem this unit, or the three years before it? Class review fixes the unit. It does not fix a prerequisite gap, and running more unit review at a student who is two grade levels behind produces compliance rather than learning. That is what Khanmigo, or a genuine intervention, is for. Getting this question wrong is the most common and most expensive error in the whole cycle.
What the research says about review that works
The evidence here is unusually clear, and it is worth knowing because it tells you which features to pay for.
The US Department of Education's Institute of Education Sciences practice guide Organizing Instruction and Study to Improve Student Learning rates "use quizzes to re-expose students to key content" at strong evidence, and "space learning over time" at moderate evidence, recommending that teachers review key content "after a delay of several weeks to several months after initial presentation" (IES practice guide). Asking deep explanatory questions also carries a strong evidence rating.
Independently, Dunlosky and colleagues, reviewing ten study techniques in Psychological Science in the Public Interest in 2013, rated only two as high utility: practice testing and distributed practice. Rereading, underlining and mnemonic devices came out low utility, with the authors noting the low-utility methods were harder to implement properly and produced inconsistent gains (Psychological Science in the Public Interest).
Our reading of that: the two methods with the strongest support are also the two that are most tedious to run by hand and cheapest to automate, which is the honest case for using any of these tools at all. It is also a warning. A tool that generates beautiful summary sheets is automating rereading, which is on the low-utility list. Pay for the quizzing and the spacing.
One legal constraint teachers are hitting this year
Worth knowing before you let AI feedback touch a grade. Four more states, Idaho, Maryland, Oklahoma and Virginia, enacted laws in 2026 requiring districts to adopt AI policies aligned with state guidance, joining Ohio and Tennessee (K-12 Dive).
Oklahoma's is the one that bites on this page: it prohibits AI from being "primarily used for grading, discipline or other high-stakes educational decisions", lets parents opt students out without academic penalty, and requires districts to disclose annually which AI tools they use and what student data those tools collect. A practice essay scored by Class Companion and recorded as a grade is exactly the shape that rule is about, and none of the seven vendors' pricing or plan pages read for this article on September 23, 2026 mentions a state constraint of this kind. Check your own state before your practice data becomes gradebook data.
On the federal side, the reason a school can put a class into most of these tools at all is the COPPA school-consent route, and that route has conditions. The Federal Trade Commission's guidance is that a school may consent on behalf of parents only where student information is used "for a school-authorized educational purpose and for no other commercial purpose", the provider gives the school the required notice, the school can review and request deletion of the data, and the provider does not use the data for targeted advertising or commercial profiles (FTC). Where a provider uses student data beyond the service the school asked for, school consent is not valid and parental consent is required.
What the directory data shows about this field
One observation from building this page, computed from the AI Educator Tools directory on September 23, 2026. Of 138 listed tools, 58 sit in an Assessment category. Of those 58, only 43 state a data hosting location and 45 state a FERPA position. So roughly one assessment tool in four tells a district nothing about where student answers are stored, in the category where student answers are the product.
That is not a reason to avoid the category. It is a reason to treat "the vendor has not said" as a question to ask rather than a gap to ignore, particularly for the tools on this page whose records are blank on all four fields. For a fuller treatment of what to check, the FERPA considerations for AI tools teachers use sets out the questions.
The bottom line
Our view: start with Wayground for the diagnostic and the drill, add Class Companion if your test is mostly extended writing, and hand students Knowt at the end. Three tools, all free at the teacher tier, covering four of the five stages. Add Conker when you need question supply, MagicSchool when reading level is the barrier, Brisk Teaching when you want to reuse what you already have, and Khanmigo when the gap is prerequisite rather than recent.
The one thing no tool on this page does is decide which two standards the class actually needs. That comes out of day one's data and your knowledge of the room, and it is the decision the whole cycle rests on. Middle school teachers looking at the same job across more subjects will find the field narrower; the AI tools that fit middle school teaching covers that, and AI tools for gifted students covers the opposite end of the same review problem.
