Privacy Policy

Last updated: 2026-10-30

This page describes, in plain language, what each edition of PointSpeak actually does with your files, your text, and your voice. It is based on reading the app's code, not on marketing language, and it covers both editions of the app separately because they are not identical.

Mac App Store edition

The Mac App Store edition of PointSpeak (also called PointSpeak Reader) only works with files and text you choose to open or send to it:

• It reads a PDF, text or Markdown file, or image that you open or paste in yourself.
• It receives selected text from another app only when you choose Read Selection with PointSpeak or Dictate into Selection with PointSpeak from that app's Services menu. When you choose Dictate into Selection with PointSpeak or Dictate at Cursor with PointSpeak and then press Replace Selection or Insert at Cursor, the text in the PointSpeak editor is sent back to that app on this Mac. PointSpeak does not look at the other app or its window.
• It recognizes text in those images, and in one window you pick through macOS's own screen-picture chooser. That single picture of the window is used to find the text in it and is not saved to disk.
• It speaks using Apple voices already installed on your Mac, including your own Personal Voice if you made one and allow PointSpeak to use it. Your Personal Voice stays on this Mac.
• It turns your speech into text using Apple's on-device Speech Recognition, and only after you press Dictate.
• It keeps the voice and speaking speed you pick, and whether you have finished or skipped the welcome, on this Mac.
• It writes technical facts about dictation and Services requests, such as character counts and how confident the speech recognizer was, to the macOS system log on this Mac. Your words, your selected text, and your audio are never written there.
• It has no account, no analytics, no advertising, and no tracking of any kind.
• The App Store edition makes no network requests. Every feature above runs entirely on this Mac, with nothing sent anywhere.

Help links open in your browser

Choosing PointSpeak Privacy Policy or PointSpeak Support in the Help menu opens this website at pointspeak-tombstone-dashs-projects.vercel.app in your default browser. Your browser connects to our hosting provider, Vercel, which can receive ordinary web-request information such as your IP address and browser details. PointSpeak does not attach your documents, selected text, or recordings to these links. These pages contain no analytics scripts.

Direct-download edition

The direct-download edition of PointSpeak includes everything the Mac App Store edition does, plus a few optional features that stay off unless you turn them on:

ElevenLabs voices (opt-in)

If you add your own ElevenLabs API key and turn that voice on, the text you choose to have read aloud is sent to ElevenLabs's servers at api.elevenlabs.io so it can be spoken in that voice. This only happens once you have supplied a key and selected an ElevenLabs voice; without a key, nothing is sent there.

A log file on this Mac

This edition keeps a plain-text log at ~/Library/Logs/PointSpeak.log so a helper who cannot hear the Mac, or a bug report, can show what happened. It can hold: every word PointSpeak has spoken aloud, each dictated transcript once Type by Voice finishes listening, and technical dictation details or error messages, such as which recognizer was used (on-device or, if you allowed it, Apple's servers), how many seconds of audio and how loud they were, and why a dictation produced no words. Each line has a timestamp. The file stays on this Mac; PointSpeak never sends it anywhere automatically. If you attach it to an email to us, we receive whatever is in it, so read it first. You can delete the file at any time.

Update checks

This edition checks for newer versions of itself at an update feed hosted at github.com. That check only learns which version is newest; it does not send your files, your text, or anything you have read or dictated.

Optional local voice bridge

If you are running a separate voice program on this same Mac, PointSpeak can talk to it at 127.0.0.1 to render speech. This stays on your Mac: it never reaches the public internet.

ElevenLabs Scribe dictation (opt-in)

Type by Voice listens on this Mac by default. If you add your own ElevenLabs API key and choose ElevenLabs Scribe under Settings, Dictation, each dictation is recorded on this Mac and sent once to ElevenLabs's servers at api.elevenlabs.io when you press the hotkey to finish, together with the words in your dictation vocabulary list, so ElevenLabs can turn it into text. Nothing is sent while you talk, and cancelling before you finish discards the recording without sending it. Turning on Never use the network in Settings turns Scribe off.

Apple speech recognition

By default Type by Voice uses Apple's on-device speech recognition and your speech stays on this Mac. On macOS 26, the first dictation asks macOS to download Apple's on-device speech model; no audio is sent for that. Settings also has a switch named Allow Apple's servers when on-device dictation gives no words. It is off by default. When you turn it on, Apple may process your speech on its servers, for a Mac whose on-device speech model is not installed yet.

Pro license

PointSpeak Pro checks whether your license is valid. This check happens offline, on this Mac, and does not contact a server.

Deleting your data and taking back permissions

Everything PointSpeak keeps stays on this Mac, so you can remove it yourself at any time.

• Mac App Store edition: quit PointSpeak, delete the app, then delete the folder ~/Library/Containers/com.tombstonedash.pointspeak.reader, which holds your saved voice, speaking speed, and welcome choice. macOS removes old system log entries on its own schedule.
• Direct-download edition: delete ~/Library/Logs/PointSpeak.log, and choose Remove next to your ElevenLabs key in Settings to delete the key from your Keychain. What ElevenLabs keeps of text or audio it already received is covered by ElevenLabs's own privacy policy.
• Permissions: to take one back, open System Settings, Privacy & Security, then Microphone, Speech Recognition, or Screen Recording (called Screen & System Audio Recording on macOS 15 and later), and turn PointSpeak off. For Personal Voice, open System Settings, Accessibility, Personal Voice.

What Tombstone Dash LLC receives

Tombstone Dash LLC, the maker of PointSpeak, receives no files, text, voice recordings, usage data, or analytics automatically from either edition of the app. If you visit our help website, its hosting provider can process the web-request information described above. If you contact us, we receive what you choose to send, such as an email.

Contact

Questions about this policy can be sent to info@tombstonedash.com.