SaaS · Collaboration · Development
Loop — unified team collaboration platform
One platform unifying real-time messaging, video meetings, scheduling, recording and task management — where communication flows naturally into action — built on a service-oriented, real-time architecture designed to scale.
The challenge
Where things stood.
Modern teams run on a growing stack of disconnected tools: chat in one app, video calls in another, the calendar somewhere else, tasks in a fourth. Every switch costs time and context.
A decision made on a call is lost because the notes live elsewhere; a request buried in a thread never becomes a tracked task; scheduling a meeting means copying details between tools. The friction is constant and invisible.
How we built it
Delivered in deliberate phases.
- 01
A real-time messaging core
Messages are handled with a persist-then-publish model — written to the database first, then published over a WebSocket layer and fanned out to connected clients through a Redis pub/sub adapter, so real-time delivery keeps working as the platform scales out horizontally.
- 02
Calling built in, not bolted on
One-to-one calls use direct WebRTC, while group video and screen sharing run through a self-managed selective-forwarding unit, and a separate media pipeline produces recordings without burdening the live call.
- 03
Communication becomes action
Any message can become an assigned, tracked task with a due date — no retyping and no separate tool — so decisions and requests made in conversation do not evaporate.
- 04
Governance for organisations
Role-based access across owners, admins, members and guests, an administration console, audit logging, instantly revocable sessions and company-wide theming and retention make the platform safe for a real organisation.
- 05
Built to scale
The application tier is stateless, with shared state held in Redis and PostgreSQL, so the platform scales by adding instances rather than being redesigned — and it was hardened through systematic architecture and security review before reaching users.
What we delivered
The work.
- Real-time messaging with channels, DMs, threads, polls, search and slash commands
- Built-in one-to-one and group video with screen sharing, host controls and server-side recording
- Tasks created directly from any message, with assignment and due dates
- Role-based access control, an admin console, audit logging and instantly revocable sessions
- A stateless application tier over Redis and PostgreSQL, engineered to scale horizontally
- A single, consistent design system with light and dark themes across the whole product
Outcome
What changed.
- Chat, video, meetings and tasks in one place, without tool sprawl
- Conversations become work — any message can be assigned and tracked to completion
- Real-time by design, on a scalable, event-driven architecture
- Engineered end to end, from data model through media pipeline to a polished interface
- A demonstration of the depth and rigour Fidova brings to a serious software product
Where most teams stitch together four tools, unifying them into one real-time product removes the everyday friction of switching — and the real engineering lives in the seams between chat, video, scheduling and tasks.
Customer support
AI-powered customer support platform
One grounded AI at the centre of every support channel — web chat, email, messaging and live phone calls — resolving routine enquiries around the clock and escalating to a person only when it should.
Sales enablement
Sales AI — real-time call intelligence
Real-time intelligence that listens to live sales calls, transcribes both sides, reads sentiment and buyer interest, and delivers in-the-moment coaching — with live quality control and performance-based lead routing for managers.
Related insights
The engineering behind the work.
Technology
The Swappable Interface: Making Build-vs-Buy a Reversible Decision
The most valuable line of code in a voice-AI system lets you replace its most important component without touching anything else.
Technology
Two processes, one system: async architecture with a background worker
Almost nothing that matters in an email-first ATS should happen inside a web request. We split the system into a web tier and a background worker with a durable queue between them, so the interface stays fast no matter what the outside world is doing.
Technology
One brain, every channel: designing a channel-agnostic support core
How we built a support platform where the channel a customer uses never changes the intelligence, knowledge or record behind it, so new channels become adapters rather than forks.