Skip to content

Audimus

Audimus is an on-device call assistant that blocks scam calls and captures your follow-ups — nothing ever leaves your phone.

Code Video Cerebral Valley

Core trick

Reads on-screen live captions instead of audio, and runs on-device Gemma to flag scam intent in the transcript.

Tags

securityproductivity

Description

Audimus is an on-device call assistant that blocks scam calls and captures your follow-ups — nothing ever leaves your phone. It reads the live call captions on-screen (no audio recorded) and reasons over the conversation with Gemma running locally via LiteRT-LM. When a call turns into a scam, it warns you in real time — reasoning about intent, not keywords. It also captures meetings and tasks and drops them into your calendar in one tap. Private, offline, local-first — powered end-to-end by Gemma.
KrishVenky/Audimus★ 00 forksKotlinMITCreated Jul 11, 2026 · last push Jul 12, 2026

README excerpt

# Audimus

An on-device AI call assistant built for the Indian context, designed to protect people when they're at their most vulnerable: on a phone call.

Built at the **Google DeepMind Bangalore Hackathon 2026**, hosted by Google DeepMind and Cerebral Valley.

## What it does

- **Real-time scam protection**: analyzes the live call transcript on-device and warns you mid-call when the conversation shows signs of a scam.
- **Follow-up capture**: detects meetings and action items mentioned during a call, then shows an editable review sheet when the call ends so you can confirm them into your calendar and tasks.
- **Privacy-first**: everything runs locally on the phone using Gemma. No audio is recorded, and nothing ever leaves the device.

## How it works

Audimus reads the on-screen transcript produced by your phone's captioning app via an accessibility service and runs it through an on-device Gemma model to classify risk and extract follow-ups. Results surface as a lightweight overlay during the call and a review sheet afterwards.

## Tech stack

- Kotlin + Jetpack Compose
- Room for local storage
- Gemma running fully on-device
- Android accessibility services for transcript capture

## Team

Built by Krishna Venkatesh and Kaushik Saravanan.

## License

[MIT](LICENSE)

Team