Verbatim includes third-party AI technologies used for transcription and language processing. This page explains how the app is designed to handle audio, imported media, transcripts, summaries, translations, and related information.
Verbatim is built around on-device processing whenever supported by the installed models and active features. For most users, this means the core recording, import, transcription, translation, summarization, and AI search workflow is intended to stay local to the device.
Last updated: August 9, 2026.
Verbatim is designed as an offline-first application for voice transcription, summarization, translation, and semantic search. The app is intended to process audio and text locally on the device whenever supported by the installed models and active features.
In normal use, Verbatim is designed so that recorded audio, transcripts, summaries, English translations, and AI-assisted search results do not need to be sent to external servers for the app’s core workflow.
Depending on how the app is used, Verbatim may handle:
Verbatim is designed so that audio recordings, imported media selected by the user, transcript content, summaries, translations, and search results remain on the device for the app’s core processing functions.
The app does not rely on cloud-based speech recognition or cloud-based AI summarization as part of its intended on-device workflow.
Verbatim does not provide recorded audio or transcript content to third parties for remote AI inference, behavioral advertising, or cloud-based profiling as part of the app’s core design.
If a user exports, copies, or separately shares transcripts, translations, summaries, or dictation output outside the app, that disclosure is controlled by the user and is outside the app’s local-only processing model.
Verbatim includes third-party AI technologies used for transcription and language processing.
Verbatim is designed to process audio and text locally on the device whenever supported by the installed models and active features.
This means users in privacy-sensitive settings such as law enforcement, fire, EMS, legal, medical, government, and field operations can work with stronger control over where their data is processed.
Verbatim can create a separate English translation for supported recordings while preserving the original transcript. The app also includes AI Search features that answer from saved transcripts and summaries, and may provide a general AI mode that chats directly with the installed on-device model.
These features are intended to run on device as part of the app’s private workflow whenever supported by the installed models and active features.
Verbatim includes an optional custom voice keyboard. Apple requires Full Access for the keyboard and the Verbatim app to exchange dictation status and completed text through a private App Group container on the device. Verbatim does not use Full Access to collect, transmit, sell, or track what the user types.
The keyboard extension does not access the microphone or run speech-recognition models. When the user starts voice dictation from the keyboard, the main Verbatim app records and processes speech locally, then returns the completed text to the keyboard for insertion into the active text field.
Recordings and generated content that the user chooses to save remain in the app’s local storage until the user deletes them or removes the app. Voice-keyboard audio is held temporarily for on-device processing and is released when keyboard mode ends.
Shared dictation state remains in the private App Group container so the keyboard can retrieve the completed result and is replaced by later keyboard sessions. Deleting a recording in Verbatim removes its locally stored audio and associated content.
Microphone permission is used for recording and voice dictation. Photo library and file access are used only when the user chooses to import media. Location permission is optional and is used only to attach location metadata to a recording. Keyboard Full Access is used for local App Group communication between the keyboard and the main app.
Users can revoke microphone, photo library, file access, location, keyboard, or Full Access permissions at any time in iOS Settings. Revoking a permission disables the features that require it but does not transmit or expose existing local data.
Verbatim stores trial and license state in the device Keychain to determine whether features remain available during the 3-day free trial or after a one-time purchase. On Mac, this includes the license key, normalized purchase email, installation identifier, activation instance identifier, cached license status, and the date of the most recent verification.
Mac purchases and license-key services are provided by Lemon Squeezy. During checkout, Lemon Squeezy processes the contact, billing, tax, and payment information needed to complete the purchase under its own privacy terms. Verbatim does not receive or store full payment-card details.
When a Mac license is activated, validated, or deactivated, Verbatim sends the license key and activation instance information to Lemon Squeezy. The app uses the purchase email returned with the license record to confirm ownership. Audio, transcripts, summaries, translations, and AI conversations are not sent to Lemon Squeezy for licensing or payment processing.
Verbatim utilizes artificial intelligence technologies to provide transcription, summarization, and language processing features. While these technologies are designed to improve efficiency and assist users, they may produce errors, omissions, or unintended results.
Outputs generated by the application, including transcriptions and summaries, are automatically produced and may not be fully accurate, complete, or contextually correct. These outputs are provided for informational and assistive purposes only.
For users in law enforcement, public safety, medical, legal, or other professional fields, AI-generated content must not be relied upon as a substitute for independent judgment, official reporting standards, established investigative procedures, clinical review, or legal analysis.
Users are solely responsible for reviewing, verifying, and validating all information before using it in reports, documentation, decision-making, or operational workflows.
Verbatim and its developers make no representations or warranties, express or implied, regarding the accuracy, reliability, or suitability of any AI-generated content.
To the fullest extent permitted by applicable law, Verbatim disclaims all liability for any damages, losses, or consequences arising from the use of, or reliance on, information generated by the application.
By using Verbatim, you acknowledge and accept the inherent limitations of artificial intelligence systems and agree that all use of the application is at your own risk.
If you have privacy questions about Verbatim, contact glen.bucy@gmail.com.