video-db/call.md

What it solves

Call.md is a desktop application that transforms meetings into "live agent loops." It solves the problem of losing critical information during calls by providing real-time AI assistance, automated transcription, and structured post-meeting summaries, reducing the need for manual note-taking and follow-up organization.

How it works

The app captures dual-channel audio (separating the user's microphone from the system audio) and screen content. This data is sent to VideoDB for real-time transcription via WebSockets. While the meeting is active, an AI agent monitors the conversation to provide contextual suggestions, track conversation metrics (like talk ratio and speaking pace), and automatically trigger Model Context Protocol (MCP) tools to fetch information. After the call, it generates a narrative summary, key points, and action items, which can be exported to Markdown or sent to automation platforms like Zapier or n8n.

Who it’s for

It is designed for professionals who conduct frequent meetings, such as sales representatives, managers, and consultants, who need real-time coaching, automated documentation, and seamless integration with their workflow automation tools.

Highlights

• Dual-Channel Transcription: Separates your voice from other participants for clearer records.
• Live AI Assist: Provides real-time suggestions on what to say or ask during a call.
• MCP Integration: Automatically detects information needs and triggers external tools via the Model Context Protocol.
• Conversation Metrics: Monitors speaking pace, talk ratio, and monologue detection in real-time.
• Post-Meeting Automation: Generates structured summaries and action items with automatic webhook exports to CRMs.
• Meeting Prep: Includes an AI-powered setup wizard and dynamic checklists based on meeting context.