Back to Work
Jishu Labs

Kiku

Voice and audio processing tool

Full-Stack Developer2024Solo project

Live Demo

Overview

Kiku (Listen) is a voice and audio processing tool that turns recordings into actionable text. Record meetings, voice notes, or upload audio files to get instant transcriptions, summaries, and extracted action items.

The Challenge

Meetings and voice notes contain valuable information that's hard to reference later. Manual transcription is tedious, and important action items get lost in long recordings.

The Solution

Integrated Deepgram for real-time transcription with GPT-4 for intelligent summarization. Auto-extracts action items, decisions, and key points from any audio content.

Impact

100+
Hours
Audio transcribed
98%
Accuracy
Transcription accuracy
2K+
Actions
Action items extracted
10+
Languages
Languages supported

Technology Stack

Frontend

Next.js 14ReactTypeScriptTailwind CSS

Audio

DeepgramWeb Audio APIMediaRecorder

AI/ML

OpenAI GPT-4WhisperSummarization

Backend

SupabasePostgreSQLStorage

Key Features

Real-time transcription
AI-powered summaries
Action item extraction
Multi-language support