Chalkie pricing at a glance
| Plan | Monthly | Yearly | Yearly works out at | What it is for |
|---|---|---|---|---|
| Free | $0 | n/a | n/a | Trying it, and light use |
| Pro | $8.99 | $79.80 | $6.65 a month | Most individual teachers |
| Max | $16.99 | $155.88 | $12.99 a month | Heavy users and long lesson series |
| Schools | Quote only | Quote only | n/a | Whole schools, linked accounts |
All figures in US dollars, read from chalkie.ai/pricing on 28 August 2026. Chalkie AI Ltd is the company behind it.
The number most pages get wrong
Chalkie’s pricing page opens on the Yearly toggle. In that state Pro shows “US$6.65 /month” and Max shows “US$12.99 /month”.
Neither of those is what you pay monthly. $6.65 is $79.80 divided by twelve, and $12.99 is $155.88 divided by twelve. If you pay month to month, the real prices are $8.99 and $16.99.
We mention it because the monthly figures do not appear anywhere in the page’s source until you click the toggle, so anywhere quoting Chalkie at “$6.65 a month” is quoting the annual rate with a monthly label on it.
What the annual discount is actually worth
Both paid plans carry a “3 months free” badge on the yearly toggle. Here is what the arithmetic says.
| Plan | 12 months at the monthly rate | Actual yearly price | You save | That is |
|---|---|---|---|---|
| Pro | $107.88 | $79.80 | $28.08 | 26.0% |
| Max | $203.88 | $155.88 | $48.00 | 23.5% |
On Pro the badge is true and slightly understated. Nine months at $8.99 would be $80.91, and Chalkie charges $79.80, so you pay $1.11 less than three months free would imply.
On Max it is marginally overstated. Nine months at $16.99 would be $152.91, and the yearly price is $155.88, so you pay $2.97 more than three months free would imply. The saving is 23.5% rather than 25%.
Neither is a big deal, and we would rather show the working than round it in either direction.
What each plan actually limits
The plans are separated by volume, not by feature category, which makes the choice unusually easy to reason about.
| Free | Pro | Max | |
|---|---|---|---|
| Resources | 5 per week | 200 per month | 400 per month |
| AI edits | 10 per week | 400 per month | 800 per month |
| Lesson series length | not included | up to 12 lessons | up to 25 lessons |
| Slides per lesson | not stated | up to 25 | up to 35 |
| Curriculum and standards alignment | no | yes | yes |
| Remove the Chalkie badge | no | yes | yes |
| YouTube videos in lessons | no | yes | yes |
| Custom themes and fonts | no | no | yes |
| Export to Google and Microsoft | yes | yes | yes |
| Card required to start | no | yes | yes |
Read the free plan limit carefully. It is 5 resources every 7 days, not 5 a month, and Chalkie’s own FAQ says those can be standalone worksheets, full lessons, or parts of a series. For a teacher building one lesson a day that is tight. For someone making a couple of worksheets a week it is a genuinely usable free tier rather than a demo.
The jump from Free to Pro is 40x the resources, from 5 a week to 200 a month. The jump from Pro to Max only doubles them. So the real decision is Free or Pro. Max is for people already hitting 200 a month, or who need lesson series longer than 12.
Things worth knowing before you pay
No card for the free plan. Chalkie states this on the pricing page, and it is uncommon enough in this category to be worth the line.
Your lessons survive cancellation. Chalkie’s FAQ says if you cancel Pro or Max you keep access to everything you already created, and only lose the paid features. That is the answer to the question most people actually have about downgrading.
Payment methods: card, Apple Pay, Google Pay, Link Pay and PayPal.
Changes take effect at the end of the current billing period, so upgrading or cancelling does not prorate mid-cycle.
Schools pricing is quote only. The Schools tier lists custom lessons, linked teacher accounts, admin controls and tracking, and centralised billing, with no figure attached. If you are buying for more than yourself, that number comes from a conversation.
What the price does not tell you
Three things the pricing page leaves open, worth settling before you commit to a year.
What counts as a “resource”. Pro allows 200 a month and Max 400, but Chalkie’s own FAQ defines a resource loosely: a standalone worksheet, a full lesson, or part of a lesson series. If a five lesson series counts as five resources rather than one, the effective allowance is very different. That distinction is not stated on the pricing page.
Whether AI edits carry over. Pro includes 400 AI edits a month and Max 800. Nothing on the page says what happens to unused edits at the end of a month, or what happens when you run out mid month. On a tool whose value is iterating on a draft, the edit allowance is more likely to bind than the resource count.
What the Schools tier costs. It lists custom lessons, linked teacher accounts, admin controls and tracking, and centralised billing, with no figure. If you are buying for a department, ask whether Schools is priced per teacher or per school, because at $16.99 a head on Max, ten teachers is already $2,038 a year and that is the number Schools has to beat.
The annual commitment is the real decision. At $79.80 for Pro you are committing to a full academic year and then some. Chalkie bills changes at the end of the current period, so an annual plan bought in September runs past the school year. If your budget is termly, monthly at $8.99 costs $28.08 more across the year and keeps you flexible.
What “free for teachers” actually means
Almost every tool in this category leads with a free tier, and they are not equivalent. The limit is where the product design shows.
The limits that are published, and therefore checkable
Chalkie states five full lessons or activity sheets every seven days. CoGrader states 100 student submissions a month. Curipod and MagicSchool both publish free plans without a headline cap on the pricing card. Verified 4 September 2026. A published cap is much easier to plan around than an unpublished one, because you can work out in advance whether it covers a normal week for you rather than discovering it in the middle of one.
Work out your real weekly volume before choosing
Count what you actually produce in a normal week: how many worksheets, how many class sets to grade, how many lesson plans. Teachers routinely underestimate this and then hit a cap on a Sunday evening. If your week is five lesson resources, a five-per-week free tier is exactly on the line and you will exceed it in any week with a cover lesson or a new unit.
Free for you may not be free for the school
A teacher signing up personally sits outside whatever approval process the school has for student data, which matters as soon as student work goes into the tool. That is not an argument against the tool; it is an argument for telling somebody. Most districts would far rather approve a tool a teacher is already using than discover it in an incident.
The commercial logic, stated plainly
Generous teacher tiers exist because teachers are the route to a district contract, which is where the revenue is. That is a perfectly reasonable model and it has a consequence worth knowing: the free tier is likely to stay usable, because it is the marketing. What is less predictable is whether the feature you come to depend on stays on the free side of the line. Do not build a department-wide process on a free tier of chalkie pricing or anything else without a fallback.
Buying at school or district level, and why teacher pilots stall
The gap between a teacher loving a tool and a district paying for it is where most education software dies, and it is a process problem rather than a product one.
Find out who actually approves before you pilot
Curriculum, IT, data protection and procurement usually all have a say, and in most districts nobody has told teachers which of them is the blocker. Ask early. A pilot that has to be re-run after the data protection review because the first one used real student work is a term lost, and it is entirely avoidable.
Price the thing you will actually deploy
Published teacher rates and district rates are different products. MagicSchool publishes Plus at $12.99 a month and quotes Enterprise with volume discounts. SchoolAI publishes tier names only. Gradescope, now on a Turnitin domain, is institution-quoted. That means a per teacher figure multiplied by staff count is a rough upper bound rather than a quote, and it is still the number worth calculating before a call, because it anchors the conversation.
Run the pilot on the thing that fails, not the thing that demos
The demo will be a well-formed lesson plan for a well-behaved topic. Your pilot should include the awkward cases: a mixed ability class, a subject with heavy notation, a student with an access arrangement, a topic where the source material is contested. How a tool handles those decides whether teachers keep using it in November.
Name who owns it after the pilot
Every tool in this category rewards an owner and punishes shared enthusiasm. Before chalkie pricing or anything else goes past a pilot, name the person whose job includes it and confirm where the hours come from. If the answer is that the keen teacher will keep doing it, the tool leaves when they do.
How to tell whether it actually helped
Education software is unusually prone to being judged on enthusiasm, because the immediate experience of a good tool is genuinely pleasant. That is not the same as it having worked.
Teacher time is the measurable win, and the honest one
Most of the real value here is hours, not outcomes. Ask two or three colleagues to record how long planning, resource creation and marking take across one normal week, before anything is introduced. It takes minutes, it is the only way the saving is ever provable, and it is unrecoverable afterwards. A tool that returns two hours a week to a teacher has earned its place without needing to claim anything about attainment.
Be sceptical of attainment claims, including your own
Attainment moves slowly, is affected by everything, and is nearly impossible to attribute to one tool inside a single year. If a vendor claims a grade improvement, ask what the comparison group was. If you are tempted to claim one internally, ask yourself the same question. Reporting a time saving you can evidence is worth more than an outcome claim you cannot.
Watch adoption in week six, not week one
Novelty carries any classroom tool through the first fortnight. The number that matters is how many teachers are still using it after half term, and whether it survived a busy assessment period. If usage collapses in the first genuinely hard week, the tool added work rather than removing it, whatever the initial reaction was.
Decide in advance what would end it
Write down before purchase what the end of the year would have to look like for you not to renew. It converts renewal from a default into a decision, which matters in a category where the per teacher cost is small enough to stop noticing.
Student data, and the questions a school has to ask before anything is switched on
This is the part of education technology where getting it wrong has consequences beyond a wasted subscription, and it is mostly answerable in writing before adoption.
Know which law you are under before you evaluate
In the United States, student education records sit under FERPA, and services aimed at under-13s bring COPPA into scope, which requires verifiable parental consent for personal information collected from children. Schools outside the US will have their own regime. The practical effect is the same: the school, not the vendor, carries the obligation, so the vendor questions exist to let you meet it.
Ask whether student work trains the model
This is the single question to get in writing. Ask whether student inputs and outputs are used to train or improve any model, including in aggregated or anonymised form, and whether that binds subprocessors too. “We do not train on your data” and “we do not train foundation models on your data” are different sentences. Get it as a contract term, because a sales assurance does not survive a change of ownership.
Ask who else touches the data
Most tools here are built on a model provider they did not write, so student work reaches a third party. Ask for the current subprocessor list, where each processes data, and how you are notified when it changes. A vendor that cannot produce that list within a day does not have one, which tells you enough about the rest of their programme.
Ask what happens at the end of a school year
Retention and deletion, stated in days rather than described as a process. What is deleted when a student leaves, when a class is archived, when the contract ends. Ask whether deletion means removed from live systems, removed from backups, or flagged. A vendor that will commit to a number in the contract is a different proposition from one that describes an approach.
The one many schools miss: inferences about a child
If a tool generates a reading level, a predicted grade, a behaviour flag or an ability estimate, that is data about a student the student never supplied. Ask whether those inferences are covered by your export, your deletion request and your access request, because a parent asking what the school holds on their child is asking about those too.
Where the figures on this page come from
Every price here was read from the vendor’s own pricing page on 4 September 2026 and carries that date, because education pricing changes with the school year and a claim without a date is not checkable.
The pages we read
Chalkie publishes free, US$6.65 and US$12.99 a month with yearly equivalents. Khanmigo publishes $44 a year. MagicSchool publishes a free plan and Plus at $12.99 a month. Curipod publishes free and from $24 a month with a 21% yearly saving. CoGrader publishes free at 100 submissions a month and $15 a month billed annually. Brisk Teaching publishes a free teacher plan at $0.
The ones we could not read
SchoolAI publishes tier names with no figures. Gradescope now sits on a Turnitin domain and is institution-quoted. We could not read a figure for Diffit, Quizlet, Photomath or Gauth either. That is not evidence they publish nothing, and we do not present it as such. Treat any figure for those six from elsewhere as unverified.
What we do not do
We do not carry a figure we cannot source to the vendor. Where a number circulates and cannot be traced, we say so and withdraw it rather than repeating it with a hedge.
How we checked this
We read Chalkie’s own pricing page in a browser on 28 August 2026, reached from the navigation on chalkie.ai rather than from a guessed URL, and we clicked the billing toggle to both states before recording anything. The monthly figures are not present in the page until the toggle is switched.
This is a pricing page, not a hands-on review. Nothing here is scored and we make no claim to have tested the product. Where figures come from Chalkie’s own FAQ rather than the pricing table, we have said so.
Prices change. If you find a figure here that no longer matches Chalkie’s page, tell us and we will correct it.
Written by
FazFaz is the founder of AIToolsBakery. Some tools here are tested hands on. Others are assessed from vendor documentation and pricing verified on the live page, and every review says which one it is. Sponsors can buy a position in a guide. They cannot buy the score, the criticism, or silence about a better option.
How we test and how we make money →Frequently Asked Questions
