Acceptable Use Policy

Bricolage AI is an autonomous research platform for legitimate professional work. Run deep research across the open web, the literature, and your own documents, and produce cited, board-ready reports, presentations, and briefs with a verifiable audit trail.

Do not use Bricolage to generate, distribute, or promote any of the following.

• Child sexual abuse material (CSAM) or content that sexualizes minors. We report violations to authorities and to NCMEC.
• Non-consensual intimate imagery, including deepfakes of real people in sexual contexts.
• Content that promotes terrorism, violent extremism, or organized hate against protected classes.
• Content that encourages or facilitates self-harm, suicide, or eating disorders directed at specific individuals.
• Material that infringes intellectual property you do not have rights to use.
• Content that defames identifiable individuals or organizations.

Do not use Bricolage agents to:

• Send spam or bulk unsolicited messages, including via Bricolage-managed email, SMS, social, Discord or any messaging integration. Comply with CASL, CAN-SPAM, GDPR, and TCPA.
• Phish, deceive, or impersonate real people or organizations without their consent.
• Build, distribute, or operate malware, ransomware, or exploits.
• Conduct credential stuffing, password spraying, or brute-force attacks against any service.
• Scrape websites, APIs, or databases in violation of their terms of service or robots.txt.
• Generate fake reviews, fake engagement, fake social proof, or astroturfed campaigns.
• Operate gambling, financial trading, lending, or insurance services without all required licenses in every jurisdiction you operate in.
• Practice medicine, law, or financial advisory services without the licenses required to do so.
• Discriminate against protected classes (race, ethnicity, national origin, religion, sex, gender identity, sexual orientation, disability, age, or veteran status) in housing, employment, credit, healthcare, or other regulated decisions.
• Make automated decisions that materially affect a person's legal rights, employment, credit, housing, healthcare, or insurance without meaningful human oversight and the disclosures required by law.
• Bypass our cost ceilings, rate limits, security controls, or approval gates.
• Resell, sublicense, or wrap Bricolage as the primary value of a competing service without a written agreement.

Outputs are generated by language models. They can be inaccurate, incomplete, or biased. You are responsible for the work your agents ship.

• Disclose to your end users when they are interacting with an AI system, where the law requires it (this includes EU AI Act Art. 50, California SB-1001, and similar rules).
• Do not rely on Bricolage outputs alone for legal, medical, financial, or safety-critical decisions. A qualified human must review.
• Do not represent Bricolage outputs as the work of a real, identifiable human without that person's consent.
• If you use Bricolage to create voice or visual content of a real person, you must have their permission.

• Do not probe, scan, or test the vulnerability of any Bricolage system or network without our written permission.
• Do not interfere with the service through denial-of-service, traffic floods, or anything that degrades availability for other users.
• Do not attempt to access another customer's data, infrastructure, or workspace.
• Do not reverse-engineer the platform to extract model weights, prompts, or proprietary techniques.

We respond to clear violations on a sliding scale: warning, feature restriction, suspension, account termination, and report to authorities for criminal activity. We may take action without notice when there is a credible risk of serious harm.

We do not refund subscription fees for accounts terminated for violations of this policy. Data export is available for thirty days after termination unless legal process requires otherwise.

If you see Bricolage being used to violate this policy, email [email protected]. Include the URL, account, or workspace you are reporting and any evidence you can share. Reports of imminent harm get a response within 24 hours.

We update this policy as the platform and the legal landscape change. The "Last updated" date at the top reflects the most recent version. Material changes will be communicated via email and posted to the changelog.