Legal
Acceptable Use
What this website and the systems we build may not be used for, including the work we decline.
Last updated 2 September 2026
Draft — pending legal review. This document was prepared by the engineering team during the build of this website. It has not been reviewed by a qualified legal adviser and should not be relied on as a statement of your rights or our obligations until it has been.
Known gaps a reviewer must close
- Whether the engagement-scope commitments below should be repeated in the master services agreement, where they would be contractually binding rather than a published position.
Using this website
This policy forms part of the Terms of Use and is published by Bralak AI Pvt. Ltd.. When using this site, do not:
- Attempt to gain unauthorised access to the site, its endpoints or its infrastructure.
- Submit another person’s personal data through the contact form without their knowledge.
- Use the contact endpoint to send bulk, automated or unsolicited messages.
- Scrape the site at a rate that degrades it for anyone else.
- Reproduce the site’s content or design and present it as your own.
Work we will not take on
This part is a statement of position rather than a rule we impose on you. We build systems that take real actions in real businesses, and some applications of that capability are ones we decline:
- Systems designed to deceive a person about whether they are speaking to a machine, where that deception is the point.
- Surveillance of individuals without a lawful basis, and monitoring of employees beyond what has been disclosed to them.
- Automated decisions about a person’s credit, employment, housing, insurance or access to healthcare taken without human review.
- Generation of content intended to impersonate a real person or organisation.
- Anything unlawful in the jurisdictions the system will operate in.
Where a person stays in the loop
Every agentic system we build classifies actions by whether they can be undone, and anything irreversible requires a person by default. Where a client asks for that threshold to be widened and the measured error rate supports it, we will.
We will not remove human review from a decision class that carries legal or safety consequence for an individual, and that is not a matter of pricing. If that is the engagement, we are the wrong supplier and will say so at the first conversation rather than the last.
Reporting a problem
If you have found a security issue on this site, or believe something we have built is being used in a way this policy prohibits, email info@bralakai.com. Please give us a reasonable period to respond before disclosing publicly.
Also