Stella is the AI operating system for accredited language schools — connecting enrolment, placement, lesson planning, assessment, and communication into one intelligent system that runs on your own hardware.
Every system in Stella feeds the next. Class records become lesson plans. Lesson plans become assessments. Assessments improve the next lesson. The loop runs overnight, every night, for every classroom.
Each module has a name and a specific job — available individually or as part of the complete Stella platform. New modules are added as the platform grows.
The heart of Stella. CLARA reads every class record, generates personalised lesson plans, and assesses progress in real time — creating the loop that makes every classroom smarter every single day.
Available 24/7 in multiple languages. Marya handles enquiries, qualifies leads, and routes prospective students to the right person — so your team focuses on decisions, not inbox management.
Front-of-house AI that greets visitors, handles walk-in enquiries, and supports students from day one. Built entirely on local hardware — no external API calls, ever.
Places students by conversation, not by multiple choice. Rachel delivers a full CEFR profile and needs analysis to teachers before day one — no external test required.
Grammar support and tutoring at the student's exact CEFR level — available in-platform for enrolled students, and by subscription for anyone in the world.
Researches contacts, drafts outreach in your organisation's voice via Stella's RAG, awaits manager approval in Zulip, sends, follows up, and escalates hot leads to the right person — automatically.
A dedicated PA instance for every manager — reads the inbox, drafts replies in that person's voice, learns their style and relationships over time, and sends nothing without explicit approval.
Because Stella runs your school's data infrastructure, it can extend into every corner of school operations — not just the academic side.
Every Stella deployment ships as a dedicated NVIDIA appliance. Plug in two cables. The device connects home. We configure it remotely. Your school never touches it again — and your student data never leaves it.
The Spark appliance handles everything locally — CLARA, Leo, Iris, Marya, Rachel — running on your own NVIDIA hardware. For tasks that need more compute (Sarah serving many concurrent subscribers, or heavy inference tasks), the appliance sends a prompt to Buongiorno's infrastructure and receives an answer back.
Think of it the way your phone uses a server for maps — the question travels, your home address doesn't. Your student records, class data, and personal information never leave the box. Ever.
Start with the module you need most. Expand when you are ready. Every tier runs on your own hardware.
Ideal for schools that want to begin with classroom personalisation before committing to a full migration.
Built for schools ready to run on AI from the moment a student enquires to the day they graduate.
For schools with infrastructure they want to keep. We connect Stella to what you already have — no rip-and-replace required.
Request a demo, ask a question, or tell us about your school. We will get back to you within one working day.