Loading…
Loading…
Lingua for Schools
Why pick? Lingua for Schools delivers both halves of a modern language program at one per-seat price. The classroom-OS half: SSO/SCIM rostering, gradebook passback to Canvas / Google Classroom / Schoology, FERPA-aligned data residency, signed COPPA + DPA. The AI-tutor half: 50 languages including Arabic + Mandarin + Korean + Hindi + Russian (Khanmigo: Spanish only), CEFR-aligned curriculum, pronunciation scoring, on-device AI via Apple Foundation Models on iPad (removes the per-seat token bill schools pay with Khanmigo or Duolingo for Schools). From $15 per seat at the 25-99-seat entry tier, $11 from 100-499 seats, $8 at 500+ seats. A 200-seat ESL program runs $2,200/mo — half what Lingoda Teams quotes for seat counts that comparable, and Lingoda doesn't even ship the AI tutor.
From 70 hours of admin a month to ~8.
School admins lose ~70 hours/month to roster sync, billing math, parent reports, compliance prep, and substitute coverage. Lingua handles each of those as a named automation. Each tutor on top gets their own ~13 hours/week back — the numbers stack.
For your team
Lesson prep auto-suggests from the engine. Parent reports auto-aggregate by family. Roster changes auto-flow into next week. The Sunday-night planning marathon turns into a 20-minute review on Monday morning — across every tutor on staff.
For you
Roster sync from Clever/ClassLink/Google Classroom. Per-seat billing with sibling discounts and scholarship awards as policy, not as email. State homeschool law compliance + FERPA + COPPA + IDEA on a single dashboard. The quarterly board prep is a 30-minute report, not a weekend.
For your growth
62 hours of operator capacity back means the second site is staff-able. The cohort builder spins up a parallel section in an afternoon. SCIM provisioning auto-creates tutor accounts when your IdP pushes new staff. The growth that always sat one hire away is now one toggle.
Week-by-week breakdown. 28 automations grouped by burden removed. The full math.
Numbers the board asks about.
Per-student daily active streaks tracked in the admin dashboard. Cohorts with assigned daily AI conversation hit 64% 30-day retention vs 22% baseline in our pilot programs.
Track active vs idle seats per cohort. Reallocate underused seats to incoming students mid-semester. No paying for seats that aren't being used.
AI-graded homework returns weekly time-on-grading per instructor. Pilot programs report 4-7 hours per week per full-time instructor freed up.
Per-tenant data residency. US-only or EU-only data on request. Signed DPA. Audit-ready data export. No third-party tracking pixels on student-facing pages.
Parents see their child's progress through a tutor-branded share link — recorded reading-aloud audio, error patterns, mastered vocabulary, time-on-task. Pilot parent NPS averaged +47 after 60 days.
The boxes your district counsel needs ticked.
Why pick AI tutor OR classroom OS? Take both.
Khanmigo gives schools a free AI tutor — for one language (Spanish), in English-speaking classrooms, funded by Microsoft Azure lock-in. Lingoda Teams gives schools a teacher-led booking platform — for 14 languages, no AI, no self-study, no on-device inference. Neither bundles both halves. Lingua does both, at one per-seat price, with 50 languages including the non-Latin scripts that Khanmigo never plans to ship.
| Feature | Koydo Lingua for Schools | Khanmigo for Districts | Lingoda Teams | Rosetta Stone Enterprise |
|---|---|---|---|---|
| Languages taught | 49 | Spanish only | 14 | 24 |
| Non-Latin scripts | Arabic, Mandarin, Korean, Hindi, Russian, Japanese, Thai, Greek | None | Mandarin only | Limited |
| Built-in AI tutor | Yes — included free, every seat | Yes — free Spanish | No | Limited / human-led |
| On-device AI (no per-seat token bill) | Yes — Apple Foundation Models on iPad | No — Azure tokens | N/A | No |
| SSO + SCIM rostering | Google, Microsoft Entra, Clever, ClassLink, SAML | Google, Microsoft | Microsoft only | SAML |
| LTI 1.3 + gradebook passback | Canvas, Google Classroom, Schoology | Khan-only | No | Limited |
| Published per-seat pricing | Yes — $15 / $11 / $8 by seat band | Free tier + $10/seat plus | Quote-only | Quote-only |
| Parent share-link (under-13 ready) | Yes — built-in COPPA flow | Limited | No | No |
Sources: vendor public pages as of May 2026. Cross-checked against Khan Academy Districts pricing, Lingoda Teams sales decks, and Rosetta Stone Enterprise reseller quotes.
From signed DPA to first student-active seat in two weeks.
A typical school pilot moves on a fortnight timeline. Week one: the curriculum director and the district IT lead sign the Lingua Data Processing Agreement (a single PDF, modeled on the CSPC standard; we will sign yours if your district has a preferred template). Lingua's solutions team enables the tenant, pre-loads the cohort with a CSV roster, and walks IT through SAML metadata exchange for whichever provider the district runs — Google Workspace, Microsoft Entra (Azure AD), Clever, or ClassLink. SCIM 2.0 keeps the roster in lockstep with the SIS thereafter, so adds/drops mid-semester sync automatically with no admin labor.
Week two: instructors land in the teacher dashboard. They pick a CEFR target per cohort (A1, A2, B1, etc.), set the language and the weekly minute commitment, and the system schedules per-student daily AI practice + weekly graded assessments. LTI 1.3 grade passback wires the gradebook to Canvas, Google Classroom, or Schoology — assessments flow back as scored, time-stamped entries instructors can override. Parents of under-13 students complete the COPPA consent flow before any audio or transcript leaves the device; over-13 students self-enroll via the SSO portal. The first measurable cohort metric — minutes-of-active-practice per student per week — is visible in the admin dashboard inside seven days of go-live.
Counsel typically asks four questions. (1) Where does student data physically live? The default is US-East AWS via Supabase Postgres; EU-only data residency is available on request at signup. (2) Who can see student audio? Only the assigned instructor, the student, and (for under-13s) the parent who consented. No third-party tracking pixels run on student-facing pages. (3) What happens if we cancel? CSV gradebook export persists at the admin level; student progress records export as a structured JSON dump per cohort. No data lock-in. (4) Who is liable if a parent sues? Lingua's DPA caps liability per the SEDTA model; KOYDO LLC carries general-liability + cyber coverage, certificate available on request from the procurement contact.
Cost discipline is the part curriculum directors most often underestimate. The school pays for active seats only — a seat goes dormant after 30 days of zero student activity and stops billing the following month. Reallocate the seat mid-semester to an incoming student with one admin click; no minimum-seat lock for the new student. Premium AI (Claude Sonnet 4.6) is a separate per-seat add-on toggled by cohort, so an AP Spanish cohort can run Premium while a freshman intro Spanish cohort stays on the included GPT-4o-mini tier. The published per-seat rate is the only line item; SSO, SCIM, LTI, video, content licensing are all included, not surcharged.
Per active seat. No quote calls.
25–99 seats
$15
per seat / mo
100–499 seats
$11
per seat / mo
500+ seats
$8
per seat / mo
25-seat floor. Annual billing = 2 months free. 30-day pilot, no card. Premium AI add-on: +$4 / +$3 / +$2 per seat respectively.
Khanmigo gives schools a free AI tutor — for Spanish only, in English-speaking schools, and the AI runs on Microsoft Azure tokens that scale with usage. Lingoda Teams and Rosetta Stone Enterprise are quote-only — every prospective school spends three sales calls to learn pricing. Lingua publishes prices ($15 per seat at 25-99 seats, $11 from 100-499 seats, $8 at 500+ seats), ships 50 languages including non-Latin scripts that Khanmigo never plans to add (Arabic, Mandarin, Korean, Hindi, Russian, Japanese, Thai, Greek), and bundles the classroom OS (SSO, SCIM, gradebook passback, FERPA-aligned data residency) at the same per-seat price. The AI tutor runs on-device via Apple Foundation Models on iPad — zero per-seat token bill for schools with iPad fleets.
25 seats. A 25-seat deployment is $375 per month or $3,750 per year. Pilot schools can run 30 days free up to 25 seats.
Yes. Google Workspace, Microsoft Entra (Azure AD), Clever, and ClassLink at launch. SAML 2.0 + SCIM 2.0 for districts. Free with every school plan; no SSO surcharge.
Yes. Native LTI 1.3 + grade passback to Canvas, Google Classroom, and Schoology. CSV export for any other LMS.
FERPA-aligned data handling baked in (US-only data residency on request). Built-in COPPA parental-consent flow for under-13 students inherited from Koydo's K-12 platform. Signed Data Processing Agreement included.
Pro Educator covers unlimited students for one tutor at a $19 flat rate. School per-seat pricing scales differently — at 25 seats, a school pays $375 per month (about 20 times the educator rate, in line with the 25-fold increase in students taught). Volume discounts kick in at 100 seats ($11/seat) and again at 500 seats ($8/seat — about half the entry rate).
Every seat includes a built-in GPT-4o-mini AI tutor for daily practice. Premium AI (Claude Sonnet 4.6 / GPT-5 tier) is an add-on: +$4/seat at 25-99 seats, +$3 at 100-499, +$2 at 500+. Toggle on or off per academic term.
Month-to-month plans cancel any time. Annual plans cancel at renewal. CSV gradebook export is permanent — no data lock-in.
25 seats free for 30 days. Signed DPA. No multi-year minimum.
Book your demo