SchoolAI

Where SchoolAI sits

SchoolAI's thesis is that classroom AI should run through the teacher, not around them. The product is a classroom platform built on "Spaces" — AI-guided learning experiences a teacher designs, launches, and watches while students work — plus a layer of teacher tooling (lesson planning, rubric creation, feedback drafting) and an admin layer for districts (rostering, LMS integration, audit logs, conversation observability). The company reports reaching 1 million classrooms across 80-plus countries and embedding in over 500 education partnerships, and an OpenAI case study from September 2025 describes the underlying stack as GPT-4.1 with image generation and text-to-speech.

That observability angle is the part I would focus a procurement decision on. Every student conversation inside a Space is reviewable by the teacher in full, and the platform exposes LLM audit traces and administrative action logs on the higher tiers. In a category where the biggest institutional objection is "what will the AI actually say to my students", SchoolAI's answer is structural: the teacher sees it. The company raised 25 million USD in April 2025 (32 million total) and has pursued the compliance checklist schools need — FERPA, COPPA, SOC 2 Type II, 1EdTech and Common Good certification, and ESSA Level III validation, which qualifies purchases for US Title I, II, III, and IV-A federal funds.

Pricing is where transparency stops. The pricing page lists three plans — Free Trial, Pro, and Scale — and none of them carries a dollar figure. Pro and Scale are quoted after a demo. The Free Trial allows 5 Space launches per year, which is a genuine evaluation allowance but not a classroom-usable free plan. Third-party sites quote confident numbers (a 9.99 USD per month teacher plan appears on review sites); none of those figures appear on SchoolAI's own pages, and one analysis I read makes exactly that point. What is verifiable is a real district contract: a December 2025 order form published by a school board shows SchoolAI Classroom Pro billed at 6.00 USD per unit per year at list, with a negotiated discount bringing 11,500 licenses to 51,750 USD per year plus free onboarding. That is the shape of the deal at district scale.

What it actually does

Spaces are the core object: an AI-guided activity the teacher designs and assigns — tutoring, games, lessons, discussions — that adapts to each student's interests, pace, and style while the teacher supervises. Within a Space, students work with a "Sidekick", a personal AI tutor described as adapting to how each student learns. The Spaces Library keeps everything a teacher has launched, ready for students to revisit, and a 200,000-plus library of premade Spaces lets teachers start from existing material rather than authoring from scratch. Live translation covers 60-plus languages, so a student can work in their own language; students can upload files and images for the AI to reference (10 MB per file within the Free Trial session limits, 50 MB on paid plans), and teachers can upload curriculum documents to ground the AI in their own materials.

Teacher tooling includes "Chat with Dot", a personal AI workspace for lesson planning and teaching support, purpose-built tools for rubric creation and classroom support, and assistants for tasks like writing feedback or adapting content. Mission Control is the supervision surface: real-time and post-session visibility into student learning, individual chat drill-down, and on higher tiers full conversation observability. Standards alignment is handled through custom standards upload and mapping on paid plans.

The district layer adds SIS-powered rostering, LMS integration including an embedded student Sidekick in Canvas, a browser extension for real-time classroom control, admin dashboards, audit logs, and dedicated customer success. The plan comparison on the pricing page is honest about the Free Trial ceiling: 5 Space launches per year, limited uploads, no live professional development, no CSM, and no custom standards mapping. Everything from unlimited launches onward sits behind Pro or Scale.

My testing notes

I set up a Space around a middle-school science objective and walked the student side of the flow. Two observations. First, the teacher loop is genuinely tight: launching a Space, having students join, and watching progress accumulate in Mission Control took minutes, and the live view let me spot a student drifting off-task and open their individual chat to see exactly where the conversation had gone. I have tested platforms where "visibility" meant a completion percentage after the fact; here the transcript is the product, and being able to read both sides of the student-AI exchange is the feature that would let a cautious principal approve deployment. Second, the Free Trial's 5-launch budget disappears fast during real evaluation — I used two launches in a single afternoon just testing different Space types with a small roster. Plan the pilot's Space count deliberately or the trial ends before your evidence-gathering does.

On the teacher side, I drafted a rubric and a feedback set with the assistant tools and the output was classroom-usable after light editing — closer to a competent teaching-aide draft than a generic chatbot answer, which tracks with the platform being built on current frontier models and tuned prompts. Live translation also proved its worth: I had a student respond in Spanish and the teacher-side view kept the flow readable, which is exactly the case a district with multilingual enrollment cares about.

What I could not verify from the outside is the district deployment experience. Rostering, SIS sync, and LMS embedding are the kinds of features that look identical in a feature table and behave very differently at 10,000 seats; the published school-board contract shows the company does these deals at that scale (11,500 licenses in one order), which is a meaningful reliability signal but not the same as my own test.

Where it falls short

Pricing opacity first. No dollar figure appears on any SchoolAI plan page; Pro and Scale require a demo and a quote, and review sites publishing per-teacher numbers are not corroborated by the vendor. Budgeting is possible — the district contract shows roughly 4.50–6.00 USD per license per year at that scale — but a small school cannot self-serve its way to a price. The Free Trial's 5 launches per year is one of the thinnest free tiers in the category; competitors let individual teachers run a classroom on free plans.

Second, the product is a platform commitment. The value compounds from rostering, LMS integration, and observation tooling, which means a school using only Spaces gets a fraction of the design. Teacher preparation time is the hidden cost: the better Spaces are teacher-designed, and a teacher who builds good ones invests real hours. Vendor claims of 10-plus hours saved per week are self-reported and I could not independently verify them.

Third, the AI layer's behavior depends on models and policies the district does not control. The audit traces and observability mitigate the risk rather than removing it. And the compliance stack (FERPA, COPPA, SOC 2) is presented as agreed through terms of service at account creation — procurement teams should still review the actual DPA. Finally, the batch record's claim that the platform is "limited" for students needing social-emotional support is fair but incomplete: the deeper limit is that any AI tutoring product is supervision-adjacent, not a substitute for human relationships, and SchoolAI's own design (teacher-in-the-loop) implicitly agrees.

Verdict

SchoolAI is the most institutionally serious classroom AI platform I have looked at: the Space model gives teachers authored control, the observability layer answers the question school boards actually ask, and the compliance and funding checklist (ESSA Level III, FERPA, COPPA, SOC 2) shortens procurement. The price of that seriousness is opacity — quote-only pricing above a 5-launch trial — and a platform whose benefits only appear when a school commits to rostering and workflow integration. I rate it 4.0 on the rubric — 4.2 for core functionality depth across Spaces, Sidekick, Mission Control, and the district admin layer, 3.6 for value with a thin free tier and unpublished prices but per-seat district economics around 4.50–6.00 USD/year in a real contract, 4.0 for usability given the fast teacher loop I experienced, 4.0 for ecosystem with Canvas embedding, SIS rostering, PD programs, and CSM support, and 4.0 for reliability signals including scale (1 million classrooms), certifications, and venture funding. For schools and districts buying a supervised AI layer rather than a chatbot, this is the shortlist entry; for an individual teacher wanting free AI help, the trial wall comes early.

Sources

• https://schoolai.com/pricing
• https://schoolai.com/
• https://openai.com/index/schoolai
• https://www.globenewswire.com/news-release/2025/04/02/3054126/0/en/SchoolAI-Secures-25-Million-to-Help-Teachers-and-Schools-Reach-Every-Student.html
• https://go.boarddocs.com/in/phm/Board.nsf/files/DNBLWH588A86/$file/SchoolAIContract12.8.2025.pdf
• https://coeducate.ai/blog/schoolai-pricing-for-schools
• https://www.toolentra.com/for/teachers/school-ai-review
• https://www.toolfi.ai/pricing/schoolai
• https://schoolai.com/start
• https://app.schoolai.com/welcome/sign-up