bolchal.ai
How It Works
Join Meeting Here
Video Conferencing Interface
Chat Functionality
AI Takeover
Return To Lobby

Bolchal.ai 👻

The Video Meeting Platform with an AI Meeting Companion

Inspiration

Modern teams rely heavily on video meetings to collaborate, brainstorm, and make decisions. However, one recurring problem we noticed was context loss.

People often:

Join meetings late
Temporarily step away from meetings
Miss important decisions or action items

When someone returns, they usually ask:

"What did I miss?"

This disrupts the flow of the meeting and wastes time repeating information.

We asked ourselves a simple question: What if meetings had a memory?

This idea led to Bolchal.ai — a video conferencing platform with a built-in AI meeting companion that listens, understands, responds when needed, and summarizes discussions so anyone can instantly catch up.

What Bolchal.ai Does

Bolchal.ai is a real-time video conferencing platform with an AI Meeting Companion that can be toggled on or off during meetings. The AI works like an invisible teammate — keeping track of everything happening so you never lose context.

Core Capabilities

Feature	Description
🎥 Real-time Video Meetings	Built-in video conferencing for seamless collaboration
🤖 AI Meeting Companion	An AI assistant that joins the meeting and listens to conversations
🔁 Temporary AI Takeover	Step away and let the AI listen, respond, and track decisions on your behalf
📝 Smart Meeting Summaries	Structured summaries of everything that happened while you were away
⚡ Instant Context Recovery	Return to any meeting fully caught up — no need to ask others to repeat

How It Works

Bolchal.ai combines real-time communication infrastructure with AI-powered conversation understanding.

Participants → Video Meeting (WebRTC)
                        ↓
             Audio / Conversation Stream
                        ↓
          AI Meeting Companion Engine
                        ↓
    Speech → Text → Context Analysis
                        ↓
   Action Items + Summary Generation
                        ↓
      User Returns → Instant Catch-Up

The AI continuously processes meeting context but only actively interacts when enabled by the user.

Technology Stack

Frontend

React / Next.js
Tailwind CSS
Real-time UI updates for live meeting state

Real-Time Communication

LiveKit for WebRTC-based video conferencing
WebSocket signaling for low-latency communication

Backend

Node.js services
Meeting session management
AI processing pipelines

AI Layer

Speech-to-text transcription
Natural language understanding
LLM-based summarization and response generation

Data Storage

Meeting transcripts
AI-generated summaries
Action items and key discussion points

The AI Meeting Companion Concept

The AI companion acts as a context-preserving layer inside meetings, modeled as:

Meeting Intelligence = Speech Recognition + Context Understanding + Knowledge Extraction

Speech Recognition — Converts meeting audio into text in real time
Context Understanding — Interprets topics, decisions, and conversation threads
Knowledge Extraction — Generates summaries and surfaces key insights

This transforms raw, unstructured conversations into structured knowledge that anyone can quickly consume.

Challenges We Faced

1. Real-Time Communication Infrastructure

Handling stable video conferencing required implementing WebRTC with low latency and reliable signaling. Integrating this alongside an AI processing pipeline — without degrading meeting performance — was a significant engineering challenge.

2. AI Context Understanding

Meetings are often non-linear. Conversations jump between topics, people interrupt each other, and decisions are made implicitly. Designing prompts and pipelines that extract meaningful summaries from messy, real-world discussions required multiple iterations.

3. Seamless AI Interaction

The AI needed to feel helpful but not intrusive. We solved this with a toggle-based assistant model — users activate the AI only when they need it, keeping it invisible during normal flow.

4. User Experience

A key UX challenge was making context recovery feel instant and intuitive — not like reviewing a transcript, but like being briefed by a smart colleague.

What We Learned

Building Bolchal.ai deepened our understanding of:

How WebRTC-based video communication platforms are architected
The challenges of real-time audio processing at scale
How to design AI systems that assist without interrupting human workflows
The value of structured, scannable AI outputs over raw transcripts

We also realized that the future of meetings isn't just better video — it's intelligent collaboration systems that help teams retain and act on information.

What's Next

Roadmap Item	Description
🤖 Q&A over meetings	Ask the AI anything about a past or ongoing meeting
📊 Meeting Analytics	Productivity insights, talk-time breakdowns, topic tracking
🔗 Integrations	Slack, Notion, Jira, and project management tools
🌍 Multilingual Support	Meeting understanding across languages
📅 Auto Task Tracking	Automatic follow-ups and action item assignment