The ground rules for using Xlaude. Written in plain language, not lawyer-speak.
The agreement between you and Xlaude.
By creating an account, installing the Xlaude app, or using the Xlaude website, you agree to these Terms of Service. If you do not agree, do not use Xlaude.
Xlaude is an AI chat and coding assistant operated by its founder, based in Qatar. "Xlaude", "we", "us", and "our" refer to the service and its operator. "You" means the person using the service.
Xlaude is a client application. It provides a chat and coding interface for AI language models that are run on your own machine or against your own API key. We do not host inference ourselves. Specifically:
Pro and Ultra are one-time purchases. You pay once and the entitlement is tied to your account forever. There is no subscription, no auto-renewal, no "lifetime access" expiring. We may launch a separate optional recurring service in the future for people who want hosted inference without BYO-key; that would be its own purchase and does not affect your one-time Pro or Ultra entitlement.
Because inference happens either on your own machine (Ollama / WebGPU) or against your own API account (BYO key), your prompts and model responses never pass through our servers. We cannot read them even if we wanted to.
The model names Spark, Spark Rogue, Ember, Ember Code, Pharos, and Pharos Rogue are Xlaude's naming for recommended underlying open-weight models (Llama, Qwen, DeepSeek, Hermes, Phi, and others). You pick which model to run through our settings; the underlying model identities may change over time as better open-weight models become available, but we will not change what each tier can access without notice.
You may use Xlaude for any lawful purpose, including creative writing, coding, research, roleplay, and conversation. Our stance is deliberately permissive: fiction is fiction, adults are adults, security research is research, and we are not your chaperone.
Unrestricted mode is a user-selectable system-prompt preset that strips the reflexive refusal patterns some underlying models have been trained with. It does not grant unlimited license. Three floor lines apply in every mode, Unrestricted included — they are baked into the system prompt itself and reinforced here contractually:
In addition to these three floors, you may not use Xlaude to:
By using Xlaude, you confirm that you are at least 18 years old (if using Pro, Ultra, or Unrestricted mode), and you accept responsibility for what you produce and what you do with it. The three floor lines above are the author of the service's choice as a matter of policy, not solely of law; they apply even where local law would allow the activity.
Before the public launch, Xlaude is available by invitation through access codes. Codes are issued at our discretion and are non-transferable. Beta access is a free preview and does not include a paid tier entitlement unless we say so in writing. We may revoke a code if it is shared publicly or used in bad faith, and we may change beta features without notice.
We aim for 99% uptime but do not guarantee it. Scheduled maintenance, third-party API outages, and events outside our control can interrupt service. Credits are considered on request for outages exceeding 24 consecutive hours (see Refund Policy).
We may change the service, add features, remove features, or update these terms. For material changes to these terms we will notify you by email at least 14 days before they take effect. Continued use after a change constitutes acceptance.
You can delete your account at any time from the account page. We may suspend or terminate accounts that violate section 5, with reasonable notice except in cases of ongoing harm. On termination, your access ends and we delete your stored data within 30 days (see Privacy Policy for details).
Xlaude is provided "as is". To the extent allowed by law, our total liability for any claim arising from your use of the service is capped at the amount you paid us in the 12 months before the claim, or $100 if you have paid us nothing. We are not liable for indirect, consequential, or incidental damages.
These terms are governed by the laws of the State of Qatar. Disputes that can't be resolved through our complaints process (see below) will be addressed in the courts of Doha, Qatar, unless local consumer protection law gives you a different forum you are entitled to.
When you get your money back.
If you are unhappy with Xlaude Pro or Ultra within 7 days of your first subscription payment, email support@xlaude.xyz with the subject line "Refund" and we will refund you in full. No reason required, no interview. We cancel your subscription at the same time.
After the first 7 days, subscription renewals are non-refundable. You can cancel at any time to prevent future renewals; your access continues until the end of the period you have already paid for.
If Xlaude is substantially unavailable for more than 24 consecutive hours in a billing period, email us and we will issue a pro-rata credit toward your next renewal.
Refunded in full, always, as soon as we see the ticket. Email support@xlaude.xyz with the charge ID.
Refunds are issued through the same payment method you used. Depending on your bank, they take 3–10 business days to appear.
If you initiate a chargeback with your card issuer before contacting us, we may suspend your account while the dispute is resolved. Please email us first — we will almost always resolve it faster.
What we collect, where it goes, how to delete it.
Your conversations are yours, and the architecture guarantees it: your prompts and model responses never pass through Xlaude's servers. On the Free tier, inference runs in your browser's WebGPU. On Pro desktop, it runs against your locally-installed Ollama. On Pro mobile, it goes directly from your browser to the inference provider whose API key you supplied, using TLS. We don't proxy any of it.
| Data | Why | Where |
|---|---|---|
| Email address | Login and account recovery | Supabase (EU region), encrypted at rest |
| Subscription status | To know which tier you're on | Our database + Lemon Squeezy |
| Chat history (Pro/Ultra, opt-in sync only) | Cross-device sync | Supabase, encrypted at rest; opt-in and client-side encrypted with your key where possible |
| Your API keys (if you supply them for mobile BYO-key mode) | So the app can call the inference provider on your behalf | Stored in your browser's local storage only — never transmitted to our servers |
| Account settings (preset, theme, model preference) | So the app opens the way you left it | Supabase if sync is on, otherwise local storage only |
When you supply your own API key on mobile, your prompts travel directly from your browser to the provider you chose. That provider's privacy policy governs that traffic. The providers we support and recommend because of their privacy commitments are:
openrouter.ai/privacygroq.com/privacyai.google.dev/termsYou are the customer of that provider; we are not in the middle. We recommend reading your chosen provider's data-retention policy and selecting one that does not train on API traffic.
Ollama is a separate open-source project (ollama.com) that you install and run locally. It is not operated by Xlaude. All inference happens on your own hardware. Nothing leaves your machine. Our website links to Ollama's installer; we do not host or modify it.
We use strictly necessary cookies to keep you signed in. No analytics or advertising cookies.
Regardless of where you live, you can:
privacy@xlaude.xyz.Xlaude is not intended for children under 13. We do not knowingly collect data from children under 13. If we learn we have, we delete it immediately.
Your data may be processed in the EU, US, and other locations where our providers operate. We rely on standard contractual clauses for transfers out of the EU.
If we suffer a data breach that affects your personal data, we will notify you by email within 72 hours of discovery, and we will tell you plainly what happened and what we are doing about it.
What's ours, what's yours, what to do if we got it wrong.
The name Xlaude, the Xlaude logotype, and the model names Spark, Spark Lite, Spark Rogue, Ember, Ember Code, Pharos, and Pharos Rogue are trademarks of Xlaude. Common-law protection applies from first use; formal registration is in progress. You may not use these marks in a way that suggests endorsement, affiliation, or sponsorship by Xlaude without written permission.
You may:
xlaude.xyz and reproduce our name in reference.Xlaude's AI capabilities are built on open-weight models released under permissive licenses by their original authors (Meta's Llama, Alibaba's Qwen, DeepSeek, Nous Research's Hermes, Microsoft's Phi, and others). We do not claim ownership of those model weights; we provide an interface and infrastructure that make them easy to use. The names Spark, Ember, and Pharos are our product-tier names and are not the underlying models' names.
You own what you write. You own what you ask Xlaude to generate. We claim no ownership of your prompts or the responses you receive.
If you believe content on Xlaude infringes your copyright, email copyright@xlaude.xyz with:
We will respond within 10 business days. We may remove content and notify the user who posted it, who has a right to counter-notify.
The Xlaude website, app, source code, documentation, and marketing materials are copyright © Xlaude. All rights reserved except as explicitly licensed.
If you supply your own API key from a third-party inference provider (OpenRouter, Groq, Google AI Studio, or any other), you are responsible for complying with that provider's terms of service and for any charges on your account with that provider. Xlaude only passes requests through; your contract for the inference service itself is with the provider. If a provider changes its terms, raises prices, or terminates your account, that is between you and them.
What we do to protect your account.
security@xlaude.xyz immediately. We can lock your account within minutes.If you find a security issue, email security@xlaude.xyz with a reproduction. We respond within 48 hours. We do not currently pay bug bounties but we will credit researchers who want public credit.
Our optional local agent (which lets Xlaude Ultra run shell commands and edit files) listens only on 127.0.0.1 (your machine's loopback interface) and authenticates every command with a per-user token stored in your OS keychain. It is open source and auditable. You can kill it with Ctrl-C at any time.
What you can turn on, off, or wipe clean.
The account page has a "Delete everything" button. One click and a 24-hour cool-off, then your account, chats, settings, and all associated records are removed. Payment records are kept for 7 years as required by tax law and reside with our payment processor.
What Xlaude actually is, and what it isn't.
Xlaude is a chat interface and coding assistant powered by large language models (LLMs). LLMs are statistical text predictors. They are useful for a wide range of tasks — writing, coding, brainstorming, explaining, searching, planning — but they do not know things in the way a person does, and they can be confidently wrong.
Code generated by Xlaude is a draft. Review it, test it, and understand it before running it, especially:
The Ultra tier's local agent asks for explicit approval before running destructive commands. Approvals are irreversible from our side — if you approve a command, we run it.
Unrestricted mode strips the refusal reflexes some underlying models have been trained with. It does not make the model smarter or more accurate. It does not grant it new abilities. It makes it more likely to answer a request fully rather than hedge or decline. You are responsible for what you do with the output.
Where any of these claims become inaccurate — if a feature changes, a provider policy shifts — we will update this page and tell affected users by email.
Who owns what you type, and what we are allowed to do with it.
You retain ownership of everything you type into Xlaude ("your prompts") and everything the AI generates in response to you ("your outputs"). Together these are "your content".
To make Xlaude work, you grant us a limited, non-exclusive, worldwide license to process your content for the sole purposes of:
This license ends when you delete your content or your account.
You are responsible for what you ask Xlaude to produce, what you share with it, and what you do with the output. Do not upload content you don't have the right to share. Do not use the service to harass or harm real people. See the Terms of Service for the complete list of prohibited uses.
If you post AI-generated text or code publicly (blog, social media, product), consider disclosing that AI tools were involved — not a legal requirement from us, but a good practice.
How to tell us something went wrong, and what happens next.
Email complaints@xlaude.xyz. Please include:
| Stage | What happens | Within |
|---|---|---|
| Acknowledgment | We confirm we received your complaint | 24 hours |
| Substantive reply | We investigate and reply with findings and proposed resolution | 7 days |
| Final decision | If a reply doesn't resolve it, we escalate internally and give you a final decision | 30 days |
If our final decision doesn't satisfy you, you can:
If you encounter content on Xlaude that violates our Terms (section 5) or that you believe is illegal, email abuse@xlaude.xyz. We take abuse reports seriously and respond within 24 hours.
Every address on this page, in one place.