Last Updated: September 24, 2026
TrialMate is used by attorneys on privileged material, so this policy is written to be specific rather than reassuring. It tells you exactly what stays on your device, exactly what leaves it, and when.
There is no account and no login. We operate no database of your cases. We collect no analytics, no usage statistics, and no advertising or tracking data of any kind, and the app contains no third party analytics or advertising software. Your trials, witnesses, notes, checklists and scanned documents stay on your device, and in your own iCloud if you turn sync on.
One thing leaves your device: the text you submit to an AI feature, at the moment you tap it. That is described in full below.
Trials and the isolation between them, witness lists and witness preparation notes, generated questions and arguments after they are returned to you, case notes, the trial checklist, and scanned documents and their extracted text are all stored on your device in the app’s own storage. Deleting the app deletes them. We have no copy and no way to retrieve them.
When you are signed in to iCloud, TrialMate automatically backs up your trials to your own private iCloud account, using Apple’s CloudKit and key value storage under your own Apple ID, so they follow you between your iPhone and your iPad and survive losing a device. That data sits in your personal iCloud account, governed by Apple’s privacy policy and your iCloud settings. We do not have access to it, and it does not pass through any system of ours. To keep a device’s trials out of iCloud, turn iCloud off for TrialMate in the Settings app, under your Apple Account, iCloud; the app then writes nothing to iCloud.
This section covers drafting witness questions, enhancing cross examination, drafting arguments, summarising documents, and generated answers on the Rules screen when you turn them on. It does not apply to any other part of the app. Searching the Rules on its own happens entirely on your device.
What is sent. When you invoke one of those features, the app sends the text of that request. It also sends the app’s bundle identifier, which identifies the application and not you. No name, no email, no account identifier, no device identifier, and no location are sent. Nothing from any other trial is ever included. Within the trial you are working in, the app sends the case material the request needs so the output is grounded in your case rather than generic. Argument generation sends that trial’s chronology, its contested facts, the witnesses tied to them, and your case notes. Drafting or enhancing questions for a witness sends the chronology entries linked to that one witness, whether each is disputed, a short excerpt of any document tied to those entries, and the questions already written for that witness. For this case context, documents are never sent in full; only the first few hundred characters of the relevant one, and the whole context block is capped, so a large matter cannot quietly send everything. Summarising a document is different: it sends that document’s full text, in sections, because a summary has to read all of it. Generated answers on the Rules screen send your question and the text of the rules it matched.
Where it goes. The request goes to an endpoint operated by ThrasherApps, which attaches the service credential and forwards the request to OpenRouter, an AI routing provider, which passes it to the language model that produces the response. The model is chosen by our endpoint, never by the app. Requests are answered by Google Gemini 2.5 Flash. If it is unavailable, a drafting request may instead be answered by OpenAI GPT-4o mini or Meta Llama 3.3 70B. Generated answers on the Rules screen are only ever answered by Gemini 2.5 Flash, and fail rather than fall back to another model. The response returns by the same path. No service credentials are stored in the app.
Retention. Our endpoint is a pass through. It does not write your request content to a database, and request and trace logging is disabled on it. Once the response is returned to your device, we hold no copy. Our endpoint instructs OpenRouter to route requests only to model providers whose published terms say they do not store submitted data or use it for training; providers that do are excluded from the route. OpenRouter and the model provider handle the request under their own published privacy terms. If retention by an outside processor matters for your matter, review their current terms before submitting anything sensitive.
Your professional obligations. You hold the duty of confidentiality, and this app does not decide for you what may be sent to an outside processor. Your jurisdiction’s rules of professional conduct, your firm policy, your engagement letters and any protective order in the matter govern that judgment. Many practitioners submit facts and issues without client identifying details. The AI features are never invoked automatically; nothing is sent unless you tap one of them.
The service and the model behind these features may change over time. Where a change alters what is described here, this policy is updated and the date above changes with it.
The camera is used only when you scan a document. Scans and their extracted text are stored on your device, and in your iCloud if sync is on. A scanned document is only transmitted if you deliberately run an AI analysis on it, in which case the section above applies. Camera access is never used for anything else.
Subscriptions are sold and processed entirely by Apple through the App Store. We never see or receive your payment details. Apple provides us only anonymous, aggregated sales reporting that cannot identify you.
When you export or share an argument set, question set, document or note, the app hands the file to the iOS share sheet and you choose the destination. That transfer happens only when you initiate it, and the app sends nothing on its own.
We do not sell, rent or trade any information. We do not build user profiles. We do not use your case content to train anything, and we have no mechanism to do so, because we do not retain it. We do not collect personal information from anyone, including children.
Deleting the app removes its local data. Turning off iCloud for TrialMate stops any further writing to iCloud, and existing iCloud data can be removed from your Apple account settings. Simply not using the AI features means no case content ever leaves your device.
We would disclose information if required by law. In practice we hold no case content to disclose, which is the point of the design described above.
If this policy changes, the date above is updated and the revised policy is posted here.
Questions about any of this? Email cbthrasher2020@gmail.com, or use the contact option in the app under Settings.
TrialMate is a preparation aid and not legal advice. See the terms of use.