Diagnosing fractions decoding ratios balancing equations

AI tutor that shows the teacher what actually changed

Students get unstuck without being handed the answer. You get a per-skill mastery record after every session, with the student's own reasoning attached.

Mastery Check

of 3 runs left

Subject

Grade band

Student work

Nothing here is stored

Use the sample or your own practice text. Never paste a real student's name.

Tutor mode

Session evidence card

This is what a teacher opens after every session. That was your third run. Nothing you typed was stored.

Student work is never used to train models. No card required to try the demo.

Not used to train models

Student work stays student work. No training, no resale, no advertising, deletable on request.

Hints before answers

Socratic mode is the default and never states the corrected value. Flip the toggle in the demo and read the difference yourself.

Evidence per skill, not a chat log

Every session ends in a mastery record you can export, present and act on tomorrow morning.

01 Why it changes the meeting

A chat transcript is not progress data

Schools bought AI tutors and got a chat box. At the end of term nobody could answer the only question that mattered. Coeducate is built backwards from that question.

Know what to reteach tomorrow

Not "the student struggled with percentages" but the actual fault line in the reasoning, plus a five-minute activity that targets it.

Not yet Successive percent change: applies both percentages to the same base.

Help that doesn't hand over the answer

The default tutor move is a question. That is what keeps the tool allowed in the classrooms that need it most, and what keeps the work worth grading.

"Twenty percent of $40 is $8. What is twenty percent of $48, and why is that a different number?"

Progress you can put in front of a board

Per-skill mastery with the student's own words as the evidence, exportable to CSV and, on District, to your SIS or LMS.

Secure Percent of a quantity, 84%, evidenced in the student's own working.

02 The session loop

Four steps, and the teacher does none of them

The evidence card is a by-product of the tutoring the student already did. No score entry, no assessment window, no transcription. See the full loop.

01

Session

A student works a problem from your material and gets stuck. Coeducate coaches with hints, never the answer.
02

Diagnosis

The engine names the specific misconception in the student's reasoning and quotes the words that show it.
03

Evidence card

Three or four skills get a state, a confidence read and an evidence line. That is the record.
04

Reteach plan

One concrete five-minute activity for tomorrow, plus a teacher note ready for a progress report.

03 A worked example

One reading answer in, one evidence card out

This is a fixed example so you can read it without running anything. The demo in the hero produces the same shape of card live, from whatever text you give it.

Input: grade 6-8 reading

Passage: The town council voted to keep the old bridge open for pedestrians only. Local business owners had asked for a full closure and a new crossing downstream. The council's report noted the repair cost was one third of the replacement cost, but did not include the annual inspection fee.

Question: What is the author suggesting about the council's decision?

Student answer. The author is saying the council decided to keep the bridge open for walking. The businesses wanted it closed. The repair was cheaper than replacing. So it was a good decision because they saved money.

Output: session evidence card

Fixed example

Diagnosis

Summarizes the text instead of inferring from it

The student restates three facts in order and then attaches an evaluation that the passage never makes. The omitted inspection fee, which is the author's actual signal, is not used at all.

Next tutoring move / hint only

The report left one cost out. Which one, and why would an author mention that a cost was left out?

Skill mastery read

Retrieving explicit detail

Secure

88%

Correctly lists the vote, the request and the cost ratio.

Inferring author stance

Not yet

18%

Attaches "it was a good decision", which the passage does not say.

Using omitted information as evidence

Not yet

24%

Does not mention the excluded annual inspection fee at all.

Citing textual evidence

Developing

47%

Paraphrases the passage but never anchors a claim to a line.

Teacher note

Retrieves detail reliably but treats inference questions as summary questions, so any "what is the author suggesting" item will fail until stance is taught explicitly.

Reteach in 5 minutes

What was left out on purpose

Read the paragraph and list every cost named.
Ask which cost is missing and who benefits from that.
Rewrite the answer starting with "The author is hinting that".

The same loop runs for math and science. See the AI math tutor, the AI reading tutor, or how a session becomes evidence.

04 For the person who signs off

Four things a curriculum director asks before signing

Evidence you can present

Per-skill mastery per student, exportable, with the student's own reasoning attached. It survives a board slide and an MTSS meeting because the wording came from the student, not from a summary.

A privacy posture built for minors

Student work is not used to train models. No advertising, no data resale, deletable on request. Designed around FERPA and COPPA obligations, and our security page states plainly which formal attestations we do and do not hold.

Rollout without a new workflow

SSO and rostering on the District tier. Teachers keep the material they already use. There is no new assessment window, because the evidence is a by-product of the tutoring session itself.

Procurement that fits your process

Invoicing and purchase orders on District and Enterprise. Named contact and an SLA on Enterprise. DPA review on Enterprise. See what each tier includes.

Check the arithmetic with your own numbers. A school on the $349/mo tier is roughly the cost of four to six hours of contracted human tutoring per month, and it covers up to 600 students. The hours you get back are the ones teachers spend working out why a student got it wrong, one paper at a time. Put your own figures into the cost-per-student estimator rather than taking ours.

05 Comparison

Where the categories actually differ

Category-level differences, not vendor claims. Where a category's answer depends on the specific plan or contract, this table says so instead of guessing. Coeducate is not affiliated with any product named here.

Capability	Coeducate	Generic AI chat	Teacher tool suites	Adaptive courseware
Student-facing tutoring session	Yes	Yes	No, the tools are for adults	Yes, inside their courseware
Per-skill mastery record for the teacher	Yes, after every session	No, only a transcript	No student session to record	Yes, within their own scope and sequence
Works over the material you already teach	Yes, paste or upload your own	Yes, but nothing is retained as evidence	Yes, for planning artifacts	No, locked to their curriculum
Hints before answers, by default	Yes, Socratic mode is the default	No, it answers on request	Not applicable	Varies by product and setting
Student work used to train models	No, never	Depends on plan and settings	Depends on vendor and plan	Depends on vendor and contract
Bought as	Monthly or annual, PO on District	Per-seat consumer or team plan	Per-teacher or district license	Multi-year district contract

Student-facing tutoring session

Coeducate: Yes
Generic AI chat: Yes
Teacher tool suites: No, the tools are for adults
Adaptive courseware: Yes, inside their courseware

Per-skill mastery record for the teacher

Coeducate: Yes, after every session
Generic AI chat: No, only a transcript
Teacher tool suites: No student session to record
Adaptive courseware: Yes, within their own scope and sequence

Works over the material you already teach

Coeducate: Yes, paste or upload your own
Generic AI chat: Yes, but nothing is retained as evidence
Teacher tool suites: Yes, for planning artifacts
Adaptive courseware: No, locked to their curriculum

Hints before answers, by default

Coeducate: Yes, Socratic mode is the default
Generic AI chat: No, it answers on request
Teacher tool suites: Not applicable
Adaptive courseware: Varies by product and setting

Student work used to train models

Coeducate: No, never
Generic AI chat: Depends on plan and settings
Teacher tool suites: Depends on vendor and plan
Adaptive courseware: Depends on vendor and contract

Bought as

Coeducate: Monthly or annual, PO on District
Generic AI chat: Per-seat consumer or team plan
Teacher tool suites: Per-teacher or district license
Adaptive courseware: Multi-year district contract

06 Pricing

Priced per program, not per seat

Annual billing is about 17% lower. There is no free tier, and the privacy posture is identical on every plan.

Classroom

$89/mo

One teacher, up to 60 students, CSV mastery export.

School

$349/mo

Up to 10 educators and 600 students, roles, SSO, scheduled reports.

District

$1,290/mo

Up to 75 educators and 5,000 students, standards mapping, SIS export, PO billing.

Full pricing and Enterprise

07 FAQ

The five questions that come up first

Still stuck on something? Email app@coeducate.ai and a person answers, usually within one business day.

Does it just give students the answers?

No. The default tutor mode is Socratic hints: the next move is a question that makes the student do the thinking, and it does not contain the corrected value or the finished step. There is a second mode, Full worked explanation, for the moment a student is genuinely stuck or a teacher wants a model answer. You can switch between the two in the demo on the homepage and read the difference in the output yourself.

Is our students' work used to train AI models?

No. Student work is not used to train models, it is not sold, and it is not used for advertising. Records are deletable on request. This is a product decision rather than a setting, which is why it is stated the same way on the pricing page and in the terms.

Are you FERPA and COPPA compliant?

The honest answer is that Coeducate is designed around those obligations rather than certified against them. School-controlled data, minimum necessary collection, no direct marketing to minors, no training on student work. What we do not have is a completed third-party audit, a compliance seal or a signed state data protection agreement in place today. The student privacy page states exactly which of those exist and which do not.

Does it work with our curriculum, or only yours?

Yours. Coeducate reads the work your students produce from the material you already teach. There is no content library you have to adopt first, which is the usual reason an AI tutoring pilot stalls in a curriculum committee.

How is this different from Khanmigo or ChatGPT?

The unit of output. A general chat product produces a transcript, and a transcript is not progress data. A content platform produces progress inside its own scope and sequence. Coeducate produces a per-skill mastery record over your material, with the student's own reasoning quoted as the evidence. The comparison table on the homepage sets out the category differences.

See what a session actually produces

Run the Mastery Check on the sample work, or paste your own practice text. Nothing you type is stored.