All work
MultiScribe logo
Case study · AI · Mobile App

MultiScribe

Record a meeting, get a speaker-labelled transcript and a summary you can send.

AI/MLSpeech to TextMobile

Overview

MultiScribe is a mobile application that turns meeting audio and video into accurate, speaker-labelled transcripts with AI-generated summaries and action items. Record in the app or upload a file, edit the notes, then share them by email.

The challenge

Meetings produce the decisions, but the record of them is whatever someone managed to type while also trying to participate. Manual notes are slow and lossy, and the transcription tools that do exist are priced for enterprises, skip summaries, or hand back an undifferentiated wall of text with no idea who said what.

How we built it

01

Capture either way

Record audio directly in the app, or upload existing MP3, WAV and MP4 files. The same pipeline handles both.

02

AI we host ourselves

Transcription, diarization, translation and summarization all run on our own inference server. No third-party AI APIs, so the audio never leaves the stack we control.

03

Transcripts that read like a conversation

Speaker diarization labels who said what, and built-in translation opens the transcript up beyond the language it was recorded in.

04

Summaries in your format

AI summaries with a selectable style, bullets or detailed prose, plus action items pulled out automatically.

05

Edit, then send

Every transcript and summary stays editable before it leaves the app, then goes out by email or share sheet.

Inside the product

Record

Capture in-app audio

Upload

MP3, WAV, MP4

Transcripts

Speaker-labelled

AI Summary

Key points, fast

Action Items

Pulled automatically

Summary Style

Bullets or detailed

Translation

Read it in your language

Library

Every meeting, saved

Edit Notes

Refine before sharing

Email & Share

Send transcript out

Secure Sign-In

Email-verified

Account Security

Password & recovery

Where it landed

  • A complete mobile experience spanning 12 feature modules
  • Recording, transcription, summarization and sharing in one flow
  • Self-hosted AI pipeline, with no per-minute third-party API costs
  • Speaker-labelled transcripts with built-in translation
  • Shipped as an installable Android APK

Tech stack

React NativeExpoNode.jsMongoDBPythonWhisper

Have a project like MultiScribe in mind?

Let's talk