Privacy Policy
Effective and last updated September 26, 2026
Scrybert is operated by Factory 226, dba Scrybert ("we," "us"), 8834 N Main St, Edgerton, WI. Questions or requests: scrybert@scrybert.com.
This policy applies to (a) account holders and app users ("you") and (b) non-user individuals recorded by a user of Scrybert ("bystanders", see the Bystanders section below). It's written in plain language and describes what the app actually does today. For a shorter, plain-language version of our data stance, see the Scrybert Promise.
Scrybert is offered solely to individuals in the United States. We do not intend to offer the service to, or knowingly accept, users located in the EU, EEA, UK, or Switzerland.
What we collect
- Audio. Recordings made through the app or a connected wearable device while a conversation is active.
- Transcripts. Text produced from that audio.
- Summaries, action items, and pivotal moments. Generated from your transcripts: a written recap, any to-dos flagged in the conversation, and notable moments Scrybert identifies (or that you flag out loud).
- Context you provide (My People, My Things). Names and relationships of people you tell Scrybert about, and things you tell it you care about, used to make summaries more accurate.
- Account data. Your profile and any custom instructions or summarizer prompts you set up.
- Approximate location, only if you turn it on. The Bunker's "Remember where you were" switch records a coarse location once at the start of a conversation and turns it into a place name. It is off unless you switch it on.
- Analytics and crash reports. Usage data and crash diagnostics collected through Google's Firebase (see "Who it's shared with," below).
Sensitive information: read this
Conversations are unpredictable. They can and do contain sensitive information: health, religious or political beliefs, relationships, employment. Scrybert's Pivotal Moments feature is specifically designed to identify notable life events, including diagnoses, funerals, and similar. We treat that as sensitive information. Health-related inferences may also be regulated as "consumer health data" under Washington and Nevada law. See our separate Consumer Health Data Privacy Policy.
Scrybert is not covered by the Health Insurance Portability and Accountability Act (HIPAA). Conversations that include personally identifiable health information are not protected by HIPAA's rules. We treat that information as sensitive under this policy, but HIPAA itself does not apply to Scrybert.
My People, My Things, and Pivotal Moments
My People and My Things are stored in a local database on your phone, and are also saved to our server as part of your account so they can restore across devices or after a reinstall. Server-side retention of this data follows the same retention setting ("Keeper") you control for the rest of your account data. See the Retention and deletion table below.
Pivotal Moments are generated on our server, returned to your device in the response, deliberately excluded from the summary file we write to server storage, and stored only in a local database on your phone.
So that the same event is not recorded twice, your phone sends the Pivotal Moments it already holds back up with each new summary request. They are used to answer that one request and are not written to our storage.
We do not retain Pivotal Moments at rest on our servers. My People and My Things are retained at rest according to your Keeper setting, described below.
How it's processed, and who it's shared with
- Transcription (our own servers). Audio is transcribed on infrastructure we operate directly, using self-hosted open-source speech recognition. We do not send your audio to a third-party transcription service, and the recording is deleted the moment the text exists.
- Summarization (OpenAI and Anthropic). Your transcript, together with My People/My Things context, is sent to AI service providers, including OpenAI and Anthropic, to generate your summary, action items, and Pivotal Moments. Under each provider's commercial/API terms, customer content is not used to train their models, and API inputs and outputs are not retained beyond what's needed to provide the service (currently up to 30 days for abuse monitoring, per each provider's stated policy).
- Read-back (our own servers). When you tap Speak, the summary text is turned into audio by software running on our own infrastructure. It is not sent anywhere else and the audio is not stored.
- Place names (OpenStreetMap). Only if you turn on "Remember where you were": your phone sends the coarse coordinates to the public Nominatim service operated by the OpenStreetMap Foundation to get a place name back. If the switch is off, no location leaves your phone because none is collected.
- Apps you connect yourself (Todoist, Google Calendar). If you connect one, the action items you choose to send are transmitted to that service, along with the conversation's one-line description. Nothing is sent until you tap send on a specific item. We hold the access tokens for those connections on our server so the connection keeps working; disconnecting in the app deletes them.
- Hosting. Our servers run on Hostinger, which stores data at our direction.
- Analytics and crash reporting (Google). We use Google Firebase for app analytics and crash reports. Crash reports can include the state of the app at the moment it failed. If Google Signals is enabled on our analytics property, that may constitute "sharing" for cross-context behavioral advertising under the CCPA, and the opt-out below applies. We do not otherwise sell your data or share it with advertisers, and we do not run ads.
We do not use your recordings, transcripts, or summaries to train any AI model, ours or anyone else's.
Accuracy
Transcripts, summaries, and Pivotal Moments are machine-generated and can contain mistakes: misheard words, wrong names, missed context, incorrect inferences about sensitive matters. Don't treat Scrybert's output as an authoritative record of what was said; when it matters, check the transcript yourself.
Recording consent and your responsibilities
Scrybert records conversations. Recording other people without their consent can be a crime and can expose you to civil liability. Some states, including California, Connecticut, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, and Washington, and arguably others, require the consent of all parties to a conversation before it may be recorded.
You are solely responsible for how, where, when, and whom you record. By using Scrybert you represent that, for every recording you make, you have obtained all consents and given all notices required by the laws of every jurisdiction where any person in the conversation is located.
The app displays a visible recording indicator, including elapsed time, while a conversation is being captured. You can disable this indicator in Settings; doing so does not reduce your own consent obligations under applicable law.
If you were recorded and are not a Scrybert user
If someone recorded you using Scrybert and you want that recording, transcript, or summary deleted, contact us at scrybert@scrybert.com with enough detail to identify it (approximate date, time, and the account holder if known). We will not build a voiceprint of you or use voice-matching to search for your audio, and we will not require you to create an account.
Retention and deletion
Everything Scrybert makes is stored on your phone. Whether a copy also lives on our server is one setting you control, called Keeper: 0 days, 30 days, 90 days, or forever. Keeper starts at 0, which does not mean "written and then deleted"; it means the file is never written at all. The server applies your setting after a conversation is processed, at startup, and every five minutes.
Audio is the exception, and it is not governed by Keeper. A recording is deleted from our server the moment it has been turned into text, every account, every setting, including when the transcription failed. It exists on our disk only for as long as it takes to read it.
| Data type | Where | Retention |
|---|---|---|
| Audio recordings | Our server, momentarily | Deleted as soon as it becomes text. Never kept, at any Keeper setting |
| Transcripts | Your device; our server only if Keeper is above 0 | Device: until you delete the conversation or the app. Server: your Keeper setting, or never written at Keeper 0 |
| Summaries / action items | Your device; our server only if Keeper is above 0 | Same as transcripts |
| To-do list | Your device | Until you clear the item or the app. A copy rides along in the account backup below when Keeper is above 0 |
| Pivotal Moments | Your device only | Until you delete them or the app. The most recent 30 are kept |
| My People / My Things / Lister / settings | Your device, and our server as one account backup file | Device: until you delete them or the app. Server: your Keeper setting, and nothing is written at Keeper 0 |
| Diagnostic copy of the model’s raw reply | Our server | One file per conversation, overwritten each time it is summarized again. Swept every five minutes on your Keeper setting - nothing survives a sweep at Keeper 0 |
| Usage log (token counts, no content) | Our server | Kept. Records how large each request was, never what it said |
| Account data | Our server | Deleted immediately when you close your account. No grace period |
| Connected-app access tokens | Our server | Until you disconnect that app |
| Analytics and crash reports | Per Firebase’s standard retention settings |
Security
We use TLS encryption in transit, key-based (no password) access to our servers, and disk encryption at rest on our hosting infrastructure. No method of transmission or storage is completely secure, and we do not promise perfect security. We deliberately avoid marketing terms like "military-grade" or "unhackable."
Children
Scrybert is for adults 18 and older. We don't knowingly allow anyone under 18 to create an account or use the service, and Scrybert is not directed at children.
Your state privacy rights
Depending on your state of residence, you may have rights to know/access, delete, correct, port, opt out of sale/share and targeted advertising, limit use of sensitive information, and be free from discrimination for exercising these rights, including under the CCPA/CPRA and comparable laws in other states. To exercise any of these rights, including opting out of the Google Analytics sharing described above, email scrybert@scrybert.com.
How to exercise your rights
Submit requests to scrybert@scrybert.com. We will verify your identity before acting and respond within the timeframe your state law requires (generally 45 days, extendable once). You may appeal a denial, or ask any other question about this policy, at the same address.
Changes to this policy
If we change this policy in a way that matters, we will update the date at the top of this page.
Contact
scrybert@scrybert.com ยท 8834 N Main St, Edgerton, WI