Acceptable use

The French version is the official one. In case of any discrepancy with this English translation, the French version prevails.

Version 1.0 — Effective date: August 9, 2026

This is a courtesy translation. The French version in French HTML version is the official version and prevails in case of any discrepancy.

This policy forms part of the Terms of Use. It applies to you, to the members of your family workspace, and to every action taken by an agent you configured or authorized.

1. Fundamental reminder: personal use only

The Service is reserved for personal and private use. Any use by or for a business, an employer, a client, an organization, or in the practice of a profession is prohibited (section 4 of the Terms of Use).

2. Prohibited uses

2.1 Illegal activities

Any activity contrary to the laws of Quebec, Canada or the jurisdiction concerned.
Fraud, phishing, identity theft, money laundering, sanctions evasion.
Market manipulation, insider trading, dissemination of false financial information, "pump and dump" schemes.
Producing, distributing or storing child sexual exploitation material, or non-consensual intimate content.
Harassment, threats, incitement to violence, hate speech.

2.2 Security and integrity

Testing, probing or attacking the Service's infrastructure or that of a third party without written authorization.
Attempting to escape your isolated environment, access another workspace, or obtain the Service's system instructions, curricula or credentials.
Denial of service, automated bulk sending, quota circumvention, creating multiple accounts to multiply free allowances.
Prompt injection intended to disable a model provider's or the Service's safeguards.
Cryptocurrency mining, hosting relay services, reselling compute capacity.

2.3 Automation and third-party accounts

Automating an account, site or API in breach of its terms of use, its robots.txt or a technical block.
Automating another person's account without their explicit authorization, or using credentials that are not yours.
Creating fake accounts, fake reviews, fake engagement or misleading content at scale.
Mass extraction ("scraping") of personal data from platforms, social networks or registries.

2.4 Communications

Sending commercial electronic messages without the consent required by Canada's Anti-Spam Legislation (SC 2010, c. 23) and without a functional unsubscribe mechanism.
Sending bulk messages, spam, or unsolicited automated communications.
Recording a conversation or communication without the consent required by law.

2.5 Third-party data and sensitive domains

Placing into the Service personal information of third parties for which a business or organization is responsible (clients, patients, students, employees, beneficiaries).
Processing health information, biometric information, third parties' social insurance numbers, judicial records, or information covered by professional secrecy.
Using the Service to make or automate decisions having a legal or significant effect on another person (hiring, dismissal, credit, housing, insurance, admission, grading, benefits).
Using the Service in any context where a failure could cause injury or death: health, personal safety, transportation, energy, critical facilities.

3. What you must do

Supervise the automations you set up and be able to stop them.
Verify outputs before relying on them.
Comply with the terms of the third-party services your agents use.
Secure your credentials and revoke compromised access without delay.
Report without delay to support@savantisai.com any security vulnerability, unauthorized access or abnormal behaviour of the Service.

4. Technical limits and fair use

The Power plan is offered with "no fixed ceiling, subject to fair use." Use that is manifestly disproportionate, abusively automated, or that degrades service for other users may be subject to rate limiting, a prior discussion, and then suspension. We always prefer notice and discussion over suspension, except in an emergency.

5. Consequences of a breach

Depending on severity and urgency: warning, rate limiting, disabling of a feature, suspension of the workspace, termination of the account, and, where the law requires it, reporting to the competent authorities.

We do not inspect the content of your conversations for surveillance purposes. Enforcement of this policy relies on aggregate technical signals, reports and legal obligations.

6. Reporting

To report abuse, a security issue or a breach of this policy: support@savantisai.com.