AI Assistant Terms
Version 2026-11-06.2 · 2026-10-14
Keep a copy. Agreement acceptance, acknowledgment of this notice and optional processing choices are recorded separately.
Purpose
These terms supplement the Terms of use and the Privacy Notice and govern the Preuve en main AI assistant. They are accepted separately, before the assistant is first used. If they conflict on the AI assistant, this document prevails.
What the assistant does, and does not do
The assistant transforms text you provide: summary, factual and neutral rewrite, translation between French and English, extraction of dates, people and amounts, and language correction. In plans for professionals, it can also produce a first draft from your instructions and summarize a document you provide.
The assistant gives no legal advice, does not choose a remedy, court or form, does not predict the outcome of a case, does not assess whether evidence is admissible, computes no time limit and files nothing. It makes no decision about you or anyone else. Every response is labelled "AI-generated, review before use".
Your responsibility
You remain the author of every document. An AI system can be wrong, leave something out or invent a fact, citation or rule: read, check and correct every response before using it, and never present a response as verified fact without checking it. Some courts require a declaration when a filed document contains AI-generated content, or human verification of every reference: check the notices of the court concerned.
Send only the information needed for the task, and only if you are allowed to share it. A professional remains responsible for professional secrecy and client authorizations. Do not use the assistant to make or justify a decision about a person (hiring, discipline, dismissal, housing, credit).
Processing of information
The text you select is sent to Cloudflare Workers AI, our processor, to produce the response; it may be processed outside Québec and Canada, including in the United States and the United Kingdom. A confirmation precedes each request. We do not keep the text sent or the response: the response is returned to your browser. We do not give ourselves permission to train models on your content.
For 12 months we keep only the usage data needed for billing and abuse prevention: account, date, feature used, number of credits and tokens, model and outcome (success or error), and a record of the confirmation given before each request, with no text. The other rules of the Privacy Notice apply, including your rights of access and deletion.
Paid access, credits and limits
The AI assistant is available only to accounts that purchased the "AI assistant" option or a plan that includes it, as shown before payment. It is not included in the free plan or the free trial.
- One credit is one request of up to about 6,000 tokens sent and 1,200 tokens received; a longer request uses more credits, up to 5 per request. The number of credits is shown before sending and is never exceeded.
- The monthly credits shown at purchase are available as soon as you subscribe, renew on the first day of each month (UTC) and do not carry over.
- A request that fails, or whose response is cut off, uses no monthly credits; it still counts toward the daily limit.
- To protect the service, a daily limit of 60 credits per account and a per-minute request limit apply.
- The assistant may be temporarily suspended for abuse, a provider outage or a safety ceiling; a suspension attributable to the service that lasts more than 7 days entitles you to a refund of the part of the paid month not provided.
Price, cancellation and refunds
The price, billing frequency and included credits are shown before payment and repeated in the contract copy emailed to you. The option can be cancelled at any time, at no charge, with the button in Account; cancellation takes effect at the end of the paid period. A subscription whose payment is overdue is cancelled immediately and the unpaid invoice is voided. If a second subscription to the option is created by mistake, it is cancelled and refunded in full. The refund rules of the Terms of use apply, including a full refund of a first payment requested within 14 days.
Prohibited uses
You may not use the assistant to fabricate or alter evidence, impersonate someone, harass, discriminate, unlawfully monitor a person, produce unlawful content, bypass limits or security measures, or attempt to extract the model or its instructions. We may suspend the assistant for an account that breaks these rules, following the procedure in the Terms of use.
Responsibility and changes
The provisions of the Terms of use on responsibility, disputes and changes apply. A change to these terms receives a new version number and, where required, a new acceptance before the next use of the assistant. Questions: equipe@preuveenmain.ca. Preuve en main (Case in Hand in the United States) · 6801, route Transcanadienne, Pointe-Claire, Québec H9R 5J2, Canada · +1 514-695-1610 · equipe@preuveenmain.ca