How is a virtual classroom platform different from a learning management system?
By what it is responsible for. A virtual classroom runs the live hour: the room, the participants and their roles, the shared material, the interaction, the attendance record, the recording. An LMS holds the course that surrounds those hours: structure, content, enrolment, assessment, and the record of the term. Most institutions need both, connected so a scheduled session appears inside the course, launches without a second login, and returns attendance and the recording to the course record. Our learning management system page covers the course side; this page covers the session.
Do you build the video infrastructure yourselves?
No, and any supplier proposing to is adding risk you do not need. Real-time media transport is a specialised, capital-intensive problem, and mature WebRTC infrastructure providers solve it better than a project team can. We build the teaching platform on top: the room model, the role and permission logic, the pedagogical interaction, the attendance and recording integration, and the connection to your course and record systems. We select and integrate the media provider based on your learner geography, recording and residency requirements, and cost profile, and we isolate that integration so the provider can be changed without rebuilding the platform.
Why not just use Zoom or Google Meet?
For many institutions you should, and we will say so. General conferencing tools are excellent at meetings, and where the requirement is genuinely a meeting with a teacher in it, a custom platform is wasted money. The case for building appears when teaching-specific behaviour matters and generic tools obstruct it: attendance that flows automatically into the student record for funding or visa compliance, seating and grouping that persists across sessions, subject-specific tools such as code execution or musical notation, single-tenant recording under strict data residency, embedded delivery inside your own product, or per-seat licensing that has become a strategic cost at your learner volume.
How are teaching roles and permissions handled?
A classroom is not a meeting of equals, and the permission model is where teaching platforms differ most from conferencing tools. We model educator, co-teacher, teaching assistant, learner, and observer as distinct roles with different rights over muting, screen sharing, whiteboard control, recording, breakout assignment, and removal. Learners typically cannot unmute others, cannot end the session, and cannot see participant contact details. Observers, whether an internal reviewer or a parent, need presence without visibility to learners in some settings and with it in others. These rules must be enforced server side, because a client-side restriction in a classroom will be discovered and circumvented within a week.
What interactive teaching features actually matter?
The ones educators use every session rather than the ones that demonstrate well. In practice that is a shared whiteboard with multi-user annotation and persistence between sessions, screen and application sharing with the ability to hand control to a learner, polls and quick checks for understanding, hand-raising and a managed question queue, chat with moderation and the ability to restrict private messaging between learners, and breakout groups the educator can move between. Beyond that, value becomes subject-specific: a coding class needs an executable editor, a language class needs recorded speaking practice, a mathematics class needs equation input. We prioritise by what your educators do rather than by feature parity with a competitor.
How do breakout groups work?
Breakouts are simple to demonstrate and hard to make dependable, so they deserve explicit design. The platform needs assignment by manual choice, random allocation, or persistent groups that survive across sessions, the ability for an educator to move between rooms and to broadcast to all of them, group-level whiteboards and shared material that the educator can review afterwards, a defined return path so nobody is stranded when the main session resumes, and reliable behaviour when a learner drops and rejoins mid-activity. That last case is where most implementations fail, and it is the one that happens most often in real classes.
How is attendance captured and reported?
This is frequently the actual reason an institution builds rather than buys. Automatic attendance is not simply presence at the start: it needs join and leave timestamps per participant, total active duration, handling of reconnections so a learner with a poor connection is not marked absent for a dropout, engagement signals where policy requires more than presence, an educator override with reason for exceptions, and transmission into the student record system as the authoritative attendance fact. Where attendance carries funding or immigration consequences, this path needs to be auditable end to end rather than a report a member of staff copies by hand.
How does recording, storage, and playback work?
Recording is a policy decision before it is a technical one, because it captures identifiable learners. We build consent capture and clear in-session indication that recording is active, configurable scope so a recording can include the educator and shared material without learner video where policy requires, and retention rules per course or institution rather than indefinite storage. Technically that means transcoding to adaptive streaming for playback on modest connections, chaptering against session events so a learner can reach the part they need, captions and transcripts for accessibility and search, and residency-compliant storage. Recordings are usually surfaced inside the course rather than in a separate archive.
How do you handle poor connections and low bandwidth?
By designing for the least favourable case rather than the demonstration case, because a class only works if every learner can participate. That means audio prioritised over video since a class survives without faces and not without sound, adaptive bitrate and resolution per participant rather than a single stream quality, graceful degradation to audio and shared material when a learner's connection cannot support video, fast rejoin that restores session state and group assignment, and asynchronous access to material and recording for learners who genuinely cannot hold a live session. We test on constrained connections and modest devices as a matter of course.
How long does a virtual classroom project take?
A focused first release covering rooms with teaching roles, shared material, whiteboard, chat, attendance capture, and recording integration is typically four to seven months including media provider integration and a pilot with real classes. The schedule is driven less by feature count than by two factors: the media provider selection and its constraints, and the depth of integration to your course and record systems. Subject-specific tooling and advanced breakout behaviour are usually second-phase work, deliberately, because they should be shaped by observed classroom use rather than specified in advance.
How do you handle learner privacy in a live classroom?
A live classroom exposes learners to each other in ways an LMS does not, which makes the privacy design part of the product rather than a policy annex. We build for camera and microphone consent with a legitimate audio-only option, name display rules where a learner's full legal name should not be visible to the class, restriction or disabling of private learner-to-learner messaging where safeguarding policy requires, no exposure of participant contact details, recording consent and indication, and complete audit of who joined which session and who accessed each recording. Where learners are minors, guardian consent and any adult-presence requirements are designed in from the start. Our information security management is certified to ISO 27001:2013 and quality management to ISO 9001:2015; we build to your safeguarding and data protection obligations rather than claiming certifications that do not exist for software suppliers.
