Terms & Conditions

By accessing or using Ail ("the Platform", "the Service") you agree to be bound by these Terms and Conditions ("Terms"). If you do not agree, do not use the Platform. These Terms constitute a legally binding agreement between you and Lahon Technologies ("we", "us", "our").

Ail is a pay-per-task AI writing platform that enables users to interact with large language AI models through a structured, domain-specific interface. The Platform offers two usage models:

• Pay-per-package: Purchase a token budget for a specific writing domain. Use the budget across unlimited back-and-forth exchanges in one conversation.
• Bring Your Own Key (BYOK): Supply your own Anthropic API key and pay a platform fee per conversation it's used in, drawn from a prepaid credit pack. Available at signup (a dedicated account type), or at any time from any account via the toggle in Dev settings — see Section 9.

You must be at least 13 years of age to use the Platform. By creating an account, you represent that you are at least 13 years old and have the legal capacity to enter into this agreement. If you are between 13 and 18, you represent that you have your parent or guardian's permission.

You must provide accurate, current, and complete information during registration. You are responsible for maintaining the confidentiality of your password and for all activities that occur under your account. Notify us immediately at tech@lahon.in if you suspect unauthorised access to your account.

We reserve the right to suspend or terminate accounts that provide false information or violate these Terms.

Each domain provides 5 free AI responses as a trial. The trial is available once per domain per user account. Trial responses cannot be transferred between accounts, extended, or converted into paid credits.

Payments: All payments are processed by Razorpay. Prices are in Indian Rupees (INR) and inclusive of applicable taxes. By completing a purchase, you authorise us to charge the stated amount.

Token budgets and credits: Token budgets and BYO conversation credits do not expire. They remain in your account until used, regardless of inactivity.

Refunds: Due to the nature of AI-generated content, refunds are generally not available once AI responses have been generated in a conversation. If you have not sent any message after purchase, you may request a refund within 24 hours by emailing sourav@lahon.in. Refund decisions are at our sole discretion. Refunds, if approved, will be credited to the original payment method within 5–7 business days.

What is a token? AI models don't read text in whole words — they break it into smaller fragments called tokens (roughly 3–4 characters of English text each). Every message you send and every reply the AI generates is measured in tokens. This is simply how usage is metered; you never need to count tokens yourself, the app tracks this for you.

Input tokens vs. output tokens: "Input" is what goes into the AI — your message, any document you upload, and the earlier conversation history sent along for context (see your Account page's conversation memory setting). "Output" is the AI's reply back to you. For the Low, Mid, and High conversation memory settings, only output tokens count against your token budget and are reflected in the credit figure shown on a response — the length of your input, an uploaded document, or how much conversation history is sent along does not itself cost you anything extra on those settings. A longer, more detailed AI response consumes more of your budget; a longer question from you does not.

Max, and large Custom windows, are billed differently: Low, Mid, and High each cap the total amount of history sent on every message, so their input size can never grow without bound. Max does not trim history at all — every message you've ever sent in that conversation is resent in full, every time, so its input size grows with the conversation itself. Custom lets you set your own window up to 200 messages, well past High's own 30-message cap, so it carries the same risk once set that high. Because of that, when Max is selected, or when Custom is set above 30 messages, input tokens (weighted by model tier, the same way output tokens already are) also count toward your credit cost and token budget, in addition to the AI's reply. Switching to Max, and saving a Custom window above that threshold, both surface this in the app before it takes effect.

Model tiers: Each package tier (Basic, Standard, Premium) runs on a specific AI model tier: Haiku (fastest, most economical), Sonnet (balanced), or Opus (most capable). Within a conversation, you may select any model tier up to the tier included in your package for each individual message.

Token budget: Every non-trial package includes a total token budget (shown as remaining tokens in the app) covering all messages in the conversation, with no limit on the number of exchanges. Each AI response deducts from this budget based on the actual output tokens generated, weighted by the model tier used relative to your package's default tier: choosing a cheaper tier than your package's default stretches your remaining budget further, while using your package's own tier deducts at the standard rate. You can never select a tier above your package's own, so this weighting only ever benefits you.

Credits — how they're calculated: The credit figure shown next to each AI response is a readable stand-in for its output-token cost, calculated from the model tier used and the response's length. Each tier has a base cost per response (Haiku: 1 credit, Sonnet: 3 credits, Opus: 10 credits — shown next to each model option when you select it), plus a bonus for longer replies charged for every additional 500 output tokens beyond the first (Haiku: +0, Sonnet: +1, Opus: +3 per 500 tokens). Credits are not a separate balance, are not sold or purchased on their own, and are not separately charged — they are computed purely for your visibility from the same tokens already being deducted from your token budget above (on Max, and on Custom windows above 30 messages, that includes the input tokens described in the previous paragraph, converted to the equivalent number of output tokens at the same rate they actually cost, before this formula is applied).

How your payment relates to tokens and credits: When you buy a package, the ₹ price you pay is for a fixed token budget on that package's model tier — a single, one-time payment with no recurring or per-message charge. There is no separate exchange rate you need to track between rupees and tokens or credits day to day: your receipt shows the ₹ you paid and the token budget it bought (see your Account page), and each response's credit figure shows what that response drew from that budget, in the terms explained above. Token budgets and BYO conversation credits do not expire.

Upgrading a package: You may upgrade an active paid conversation to a higher tier within the same domain at any time. Upgrading keeps the same conversation and its full history, it does not start a new conversation. You are charged only the price difference between your current and new tier, discounted by 30%.

Converting a trial: Converting a free trial conversation into a paid package likewise keeps the same conversation and its history. Since no payment has been made on a trial, this conversion is charged at the package's full price.

Token topups: If you exhaust a package's token budget, you may purchase additional tokens for that same conversation without upgrading tiers or starting a new conversation. Topup token amounts are calculated based on the model tier your conversation runs on, so the tokens you receive reflect a fair value for the price paid regardless of tier.

What it is: You may optionally connect your own retrieval endpoint in Dev settings. When enabled, we send your endpoint a small request containing your message and the conversation's domain on every message, and if it responds with content, we add that content to the AI's context before it replies.

You are responsible for your endpoint: We do not review, curate, verify, or store what your endpoint returns. You are solely responsible for the accuracy, legality, and safety of any content it provides, and for your endpoint's own security, uptime, and licensing of whatever it retrieves from. Content returned by your endpoint is treated as untrusted, user-supplied input, the same as an uploaded document.

Basic safety checks, not a guarantee: We require your endpoint to be an HTTPS URL that resolves to a public address, and we apply reasonable technical safeguards before calling it. This is a baseline check, not a security guarantee of your endpoint or its infrastructure. We may disable this feature on your account without notice if your endpoint is found to return harmful content or is used to probe or attack our infrastructure.

Billing: Content returned by your endpoint is billed as input tokens, weighted by model tier the same way output tokens already are, in addition to the AI's reply — the same treatment described for the Max conversation memory setting in Section 7. This applies whenever your endpoint contributes content, regardless of your conversation memory setting.

Reliability: If your endpoint is unreachable, slow, or returns something we can't use, we proceed without it rather than failing your request — you won't be charged extra for a request where your endpoint didn't successfully contribute content.

What it is: From Dev settings, you may connect your own Anthropic API key and turn it on at any time, for any account — you don't need to have signed up as a BYOK account to use this. While enabled, your key pays for AI generation directly, and Anthropic bills you (or whoever the key belongs to) for that usage separately from anything you pay us.

Overrides normal billing: When your key is active for a message, that message does not draw from a paid conversation's token budget at all. Instead, activating your key in a given conversation requires a one-time flat activation fee, paid via Razorpay directly from that conversation the first time you try to send with your key on — further messages in that same conversation are free of additional platform fees for as long as it stays active there. Turning your key off returns that conversation to its normal package billing for later messages.

Payment required before sending: If a conversation hasn't been activated yet, sending with your key on is blocked until the activation payment completes — we do not silently fall back to platform billing or let a message through unpaid. Activation is a one-time Razorpay checkout, not a subscription or a prepaid credit pack; each new conversation you want to use your own key in needs its own activation payment.

Your responsibility: You are responsible for the validity, security, and usage costs of any API key you connect, and for complying with Anthropic's own terms of service. We store it encrypted and use it solely to generate your responses; we do not use it for any other purpose.

You agree not to use the Platform to:

• Generate content that is illegal, harmful, threatening, abusive, harassing, defamatory, or obscene
• Create malware, ransomware, phishing materials, or any other malicious content
• Generate content that facilitates violence, terrorism, or discrimination
• Produce child sexual abuse material (CSAM) or any content that exploits minors
• Infringe the intellectual property rights of any person
• Attempt to reverse engineer, disassemble, or probe the Platform's AI systems
• Use the Platform in a manner that overloads or disrupts our infrastructure
• Create multiple accounts to circumvent trial or rate limits
• Configure a custom retrieval endpoint (Section 8) to attack, probe, or gain unauthorised access to our or any third party's infrastructure

Violation of the acceptable use policy may result in immediate account suspension and, where appropriate, reporting to relevant authorities.

Accuracy: AI-generated content may contain errors, inaccuracies, or outdated information. You are solely responsible for reviewing, verifying, and editing AI-generated content before using it for any purpose.

Professional advice: Content generated in the Legal, Medical, or Financial domains is for informational purposes only and does not constitute professional legal, medical, or financial advice. Always consult a qualified professional for such matters.

Ownership: Content you generate using the Platform is yours. We do not claim ownership of your AI-generated outputs.

The Ail name, logo, platform design, and all underlying software are the intellectual property of Lahon Technologies. You may not copy, reproduce, distribute, or create derivative works from our platform without our express written permission.

Our collection and use of personal data is governed by our Privacy Policy, which is incorporated into these Terms by reference. By using the Platform, you consent to the data practices described in the Privacy Policy.

To the maximum extent permitted by applicable law, Lahon Technologies shall not be liable for any indirect, incidental, special, consequential, or punitive damages arising from your use of or inability to use the Platform, including but not limited to loss of profits, data, or goodwill.

Our total liability to you for any claim arising from use of the Platform shall not exceed the amount you paid us in the 3 months preceding the claim.

The Platform is provided "as is" and "as available" without any warranty of any kind, express or implied. We do not warrant that the Platform will be uninterrupted, error-free, or that AI responses will be accurate, complete, or suitable for any particular purpose.

We reserve the right to modify, suspend, or discontinue the Platform (or any part of it) at any time with or without notice. We shall not be liable to you or any third party for any modification, suspension, or discontinuation of the Service. In the event of permanent discontinuation, any unused credits will be refunded pro-rata.

These Terms are governed by the laws of India. Any dispute arising out of or in connection with these Terms shall be subject to the exclusive jurisdiction of the courts of India.

We may update these Terms from time to time. Material changes will be communicated via email or an in-platform notice at least 7 days before they take effect. Your continued use of the Platform after the effective date constitutes your acceptance of the revised Terms.

For questions about these Terms:

• General: sourav@lahon.in
• Technical: tech@lahon.in
• Business: business@lahon.in