Acceptable Use Policy

This Acceptable Use Policy ("AUP") governs the access and use of the website, applications, APIs, SDKs, browser extensions, and developer portals provided by Xarwiz Technologies LLP, Noida, Uttar Pradesh, India - 201301.

AI Model/LLM: Machine learning algorithms processing prompts. Prompt: Text, query, dataset, or file uploaded to generate a response. Output: AI-generated responses. Malicious Activity: Attempts to bypass system security, rate limits, or RLS parameters.

Every user must comply with applicable Indian and global laws, respect third-party intellectual property under the Copyright Act 1957, protect personal data privacy, and safeguard login credentials and API keys.

Permitted activities include business productivity enhancement, programming help, educational research, automatic database indexing, and workspace document classification using RAG configurations.

Prohibited activities include: illegal activity & fraud; harassment & hate speech; cybersecurity exploits (ransomware, keyloggers, SQL injections, CSRF attacks); adversarial AI attacks (jailbreaking, safety bypass); and unauthorized automated scraping.

Users must not use AI outputs to train competing machine learning models, deploy AI agents to generate spam or fake reviews, create deepfakes to impersonate real persons, or use AI for academic cheating.

The platform must not be the sole decision-maker for critical cases. Prohibited high-risk activities include medical diagnostics, legal advice, financial investment decisions, military operations, and nuclear system control.

Users must not upload copyrighted files or pirated software without ownership rights, misuse trademarks, or bypass digital rights management (DRM) protections.

You are solely responsible for all prompts, files, and vector indices uploaded to your tenant workspace. Ensure that uploads do not contain malicious scripts or infringe on privacy.

Do not upload proprietary source code without authorization, plaintext passwords, private key files, or sensitive personal records without active consent.

You must limit queries to the rate boundaries of your subscription tier. Prevent API key exposure (e.g. do not check keys into git), and do not deploy bots to create multiple API accounts.

Workspace Administrators are responsible for revoking access for inactive team members, enforcing Multi-Factor Authentication (MFA), and auditing workspace prompt metadata logs.

Clients must access services through updated browsers. Security vulnerabilities must be reported immediately to security@xarwiz.com. Do not exploit logical security flaws.

We run automated moderation scans on prompts to identify prohibited content, and monitor request frequencies to block mass automation, bot calls, and credential abuse.

Violations will result in formal warning warnings, temporary workspace rate limiting, permanent account termination, or reporting to legal authorities (CERT-In).

Submit reports of AUP infractions to: support@xarwiz.com (support issues), security@xarwiz.com (security concerns), or legal@xarwiz.com (copyright alerts).

You can submit review requests to appeals@xarwiz.com within 14 business days. All decisions by our trust and safety audit team are final.

We reserve the right to modify this AUP. Changes are published on this page with updated version histories.

Company: Xarwiz Technologies LLP, Sector 62, Noida, Uttar Pradesh, India - 201301. Grievance Officer: Ashish Pratap Singh Tomar (grievance@xarwiz.com).

Creating client summaries from parsed transcripts, converting legacy code to Python, or draft layout suggestions for interfaces.

Generating phishing email formats, or configuring bots to harvest database tables.

Exploiting LLMs to write malware, or generating deepfake voice profiles to authorize financial transactions.

Set strong passwords, enable Multi-Factor Authentication (MFA), and store API keys in environment variables.

Users must align query practices with: Fairness (avoid bias), Transparency (state where AI is utilized), and Human Oversight (verify critical outputs before use).