Dashboard
Voice agent
Usage Data by app
Session stats
Logo

Mira — Intelligent Work Companion

🌟 Inspiration

We noticed a pattern in our own work, endless hours in front of screens, pretending to be productive, yet unclear about what we actually accomplished.
Most productivity tools rely on time-blocking or to-do lists, but they rarely reflect reality. We wanted something that could see what we’re doing, understand the context, and help us think better, not just work harder.

That idea became Mira, an intelligent work companion that gives users a semantic picture of their day and contextual assistance, all while keeping data local.

💡 What It Does

Mira quietly runs in the background, observing digital activity to create a semantic timeline of your workday.
It recognizes:

Which apps you’re using
How long you stay focused
Captures screenshots for quick visual recall

Its interactive AI companion can be triggered with a global shortcut and operates in two modes:

🎙️ Speech-to-Speech

Talk to Mira naturally, ask what you worked on earlier, get summaries of your focus time, or ask for help when you’re stuck.

💬 Text-to-Text

Mira scans your current screen, understands context, and provides relevant insights in real time.
Example: “You seem to be debugging in React, here’s what might be wrong.”

It’s like having an intelligent coworker who knows your workflow.

🏗️ How We Built It

Layer	Stack	Description
Agent	`C++`	Native logging agent capturing app name, window title, timestamp, and idle time every second.
Context Layer	`Python`	Handles screenshot capture and semantic extraction from the visible screen for higher accuracy.
Frontend	`React + Electron + shadcn/ui`	Displays interactive charts and daily timelines
Companion Layer	`ElevenLabs + Local Context`	Global shortcut toggles AI assistant between text and voice modes.

⚙️ Key Features

🖥️ Local-first data logging (no cloud storage)
🧩 Cross-language integration (C++ ↔ Python ↔ React)
🔊 Real-time speech + text assistant
📊 Clean and modern dashboard

🚧 Challenges We Faced

Cross-platform APIs: Handling window detection differently on Windows vs macOS.
Speech synchronization: Managing real-time mic input with rendering loops.
UI design: Creating an intelligent assistant that feels helpful, not intrusive.