Technology

Built, tested and ready to adapt.

Our technology has been developed through real educational deployments. Each component can serve a specific need or become part of a larger learning system.

How the system fits together
  1. Partner programmes and approved resources
  2. Luma Content Factory and education intelligence layer
  3. Messaging, web applications and offline delivery
  4. Learners, teachers, parents and programme teams
Reporting and improvement
A · Messaging platformsDeployed

Learning through the channels people already use.

WhatsApp and Facebook Messenger experiences that deliver education on familiar, low-data channels. Each can be branded and configured for an individual organisation.

  • Curriculum-based tutoring
  • Lesson support
  • Content distribution
  • Assessments
  • Reminders
  • Programme communication
Luma Learn on WhatsApp listing Sesotho reading documents and sending the chosen PDF
Requesting Sesotho reading material through Luma on WhatsApp.
B · Lightweight web applications

Mobile-first, without the complexity.

Web applications that load quickly on modest phones and connections, for learners, teachers and the teams who run programmes.

  • Learner portals
  • Digital learning pathways
  • Assessment experiences
  • Teaching resources
  • Administrator interfaces
C · School in a BoxIn development

Learning infrastructure that doesn't wait for a signal.

A locally deployed education system for schools and communities with unreliable connectivity. It serves content and AI support on site, and synchronises when internet access becomes available.

Local deployment
Runs on a local server inside the school or centre.
Offline access
Nearby devices connect over local Wi-Fi.
Local AI
Learning support from models running on the device.
Periodic updates
New content and improvements sync when a connection is available.
Illustration of School in a Box: a solar panel powering a local server that broadcasts Wi-Fi to nearby phones
D · Content Factory

From trusted resources to digital learning.

A system for converting existing educational material into structured resources that work across channels. Human review sits at the centre: nothing reaches learners without approval from the people responsible for the content.

  1. 01

    Educational content

    Lesson plans, teacher guides, past papers, readers.

  2. 02

    Structured learning resources

    Quizzes, assessments, pathways and translations.

  3. 03

    Review and approval

    Your subject experts check and sign off.

  4. 04

    Delivery across platforms

    Messaging, web and offline from one source.

E · Education-specific AIDeployed

Not a generic chatbot inside a learning app.

Our education intelligence layer works from approved instructional and curriculum resources. It retrieves from those sources, follows structured learning pathways and is evaluated against educational expectations.

No AI system is free of error. Safeguards, evaluation and content boundaries reduce the risk of inappropriate or off-curriculum responses, and we keep improving them with use.

  1. 01

    Curriculum alignment

    Responses are anchored to the curriculum and grade in use.

  2. 02

    Approved-resource retrieval

    Answers draw on your materials before anything else.

  3. 03

    Content boundaries

    Clear limits on topics and tone for the audience.

  4. 04

    Evaluation and safeguards

    Ongoing testing of response quality and appropriateness.

F · Local and small language models

AI that runs where the learning happens.

We are developing AI capabilities that run on compatible phones, computers and local servers. Keeping processing on the device matters where connectivity is limited and where data should stay local.

G · Integration and reporting

Separate components. One system.

The products share configuration, content and reporting, so a programme can combine channels without running separate platforms.

See how these pieces fit your programme.

Build with Luma Ways to work with us