Your voice, your data. Your computer.

Talkr has no servers, so this is a short one.

Our commitment

Talkr is built on one rule: your voice, your data, your computer. We do not collect, transmit or store your personal data on any server. We don't run one.

This page explains what Talkr handles, where it lives, and what you can do with it.

What Talkr handles

Processing: All audio and text is processed on your computer. Nothing is sent to us or to anyone else. Once a model is downloaded, Talkr works fully offline.
Storage: Models, history and settings live in ~/.talkr on your machine. Delete them whenever you like. There is no cloud sync and no account.
Telemetry: None. No analytics, no crash reports, no usage statistics. Talkr goes online only to download a model you pick and to check for updates, which you can turn off.
Source: The code is MIT licensed. You can audit it, build it yourself or contribute at github.com/alihassan-coder/Talkr.
Model licenses: Whisper (MIT), Kokoro (Apache 2.0), Piper (MIT, with espeak-ng data under GPL 3.0). Each license is shown in the app before you download the model.

Where your data lives

Everything Talkr keeps sits in a single folder on your computer.

~/.talkr/

Linux and macOS

C:\Users\You\.talkr\

Windows

$TALKR_HOME

Custom location

Inside it you will find models/ for downloaded models, audio/ for generated speech, history/ for the SQLite database, cache/ for temporary downloads and logs/ for app logs.

Your rights

Access. Everything is in ~/.talkr. Inspect it, copy it or move it at any time.
Deletion. Delete the folder and every trace is gone. Or use “Delete all history” in Settings.
Portability. Export transcripts as text, Markdown, SRT, WebVTT, CSV or JSON, and speech as MP3, FLAC or WAV. Models are standard GGML and ONNX files.
No profiling. We do not build profiles, show ads or sell data. There is no account to delete.

Network connections

Talkr only connects to the internet in two cases: when you click Download on a model, which fetches it from Hugging Face or GitHub Releases, and when it checks for a new version.

The update check runs shortly after launch and then every six hours. It downloads one small file describing the latest release from GitHub Releases and sends nothing about you, your audio or your text. Like any download, GitHub sees your IP address. Turn it off with Check for updates automatically in Settings, and use Check now when you want to.

There is no analytics and no crash reporting. Any network monitor will show you the same.

Third-party components

Talkr bundles a few open-source projects. All of them run locally.

whisper.cpp: MIT. Speech recognition, from ggerganov/whisper.cpp.
sherpa-onnx: Apache 2.0. Text-to-speech, from k2-fsa/sherpa-onnx.
Tauri: MIT or Apache 2.0. The desktop framework.
React and Rust: The interface and the core runtime.

Contact

Questions about this policy? Open an issue on GitHub.