1. About These Terms

1.1 These AI Gateway Terms ("AI Terms") apply to your use of the n8nClouds AI Gateway. They supplement our Terms & Conditions, Privacy Policy and Refund Policy.

1.2 Where these AI Terms conflict with those documents, these AI Terms prevail for the AI Gateway only. Everything else in those documents continues to apply.

1.3 By creating an API key, purchasing AI Credit, or sending a request to the AI Gateway, you accept these AI Terms.

2. Definitions

"AI Gateway" — our API service at the base URL shown in your dashboard, which gives access to AI models through an interface compatible with the OpenAI chat completions API.
"AI Credit" — the prepaid balance that pays for AI Gateway usage.
"API Key" — a secret key beginning sk-bkr- that authenticates requests to the AI Gateway.
"Model" — an AI model listed as available in the AI Gateway.
"Model Provider" — the third party that operates a Model, currently OpenAI.
"Request" — a single call to the AI Gateway.
"Input" — everything you send in a Request, including text, images and tool definitions.
"Output" — everything the AI Gateway returns in response.
"Content" — Input and Output together.
"Token" — the unit Models use to measure text and images, and the unit we charge in.

3. The AI Gateway

3.1 What it provides. The AI Gateway supports:

chat completions, streaming and non-streaming;
image input (sending images as part of a message);
tool and function calling;
JSON and structured (schema) output;
reasoning models; and
listing available Models and checking your balance.

3.2 What it does not provide. The AI Gateway does not offer embeddings, audio (speech or transcription), image generation, moderation, realtime, or the legacy completions endpoint. Requests to these return an error and are not charged.

3.3 Compatibility. The AI Gateway is compatible with the OpenAI chat completions API for the features listed in 3.1. We do not promise compatibility with any other feature, endpoint, or future change to that API.

3.4 Access. AI Gateway access must be enabled on your account. We may decline or withdraw access at our discretion.

3.5 Models. We decide which Models are available. We may add, change, rename or retire Models at any time. Where practical, we will give notice before retiring a Model that is in active use.

3.6 No affiliation. The AI Gateway is our own service. We are not affiliated with, authorised by, or endorsed by any Model Provider.

4. API Keys

4.1 Each API Key is shown in full once, when it is created. We store only a protected form of it and cannot show it to you again.

4.2 You are responsible for keeping your API Keys secret and for all usage and charges made with them, including usage by anyone who obtains a key from you or through your systems, until you revoke the key.

4.3 Revoking an API Key is permanent. A revoked key cannot be restored; create a new one instead.

4.4 You must not sell, rent, share, publish or otherwise give access to your API Keys to third parties. Building your own application on the AI Gateway is allowed (see Section 10); giving others direct access to your keys is not.

4.5 Tell us immediately at support@n8nclouds.com if you believe an API Key has been compromised. Revoke it from your dashboard first.

5. AI Credit

5.1 Purchase. You buy AI Credit through an invoice payable in Bangladeshi Taka or another currency we accept. Before you pay, we show how much AI Credit you will receive. That amount is fixed at purchase and is not affected by later exchange-rate movements. Minimum and maximum purchase amounts are shown at checkout.

5.2 Separate balance. AI Credit is separate from your main account balance. AI Credit can be used only for the AI Gateway, and cannot pay for hosting or any other service. Your main account balance cannot pay for AI Gateway usage.

5.3 No expiry. AI Credit does not expire while your account is active.

5.4 Not money. AI Credit is a prepaid service balance. It earns no interest and cannot be transferred to another account, exchanged, or withdrawn as cash, except as described in Section 8.

6. Pricing and Charges

6.1 Prices. Each Model has its own price per million Tokens, shown in your dashboard. A Model may have separate prices for input Tokens, cached input Tokens and output Tokens.

6.2 What is counted. A Request is charged for:

input Tokens — the text, tool definitions and images you send. Images are converted to Tokens by the Model Provider; a single image can count as several thousand Tokens, depending on the Model and the image;
cached input Tokens, where the Model Provider reports them, at the cached rate; and
output Tokens — everything the Model generates, including any internal reasoning Tokens a reasoning Model uses before answering.

Token counts come from the Model Provider's report for that Request.

6.3 Minimum charge. Each completed Request is charged at least the Model's minimum charge, even if its Token cost is lower.

6.4 When you are not charged. You are not charged for a Request that:

we refuse before sending it to a Model (for example, an unknown Model, an invalid request, insufficient AI Credit, or a rate, capacity or spending limit);
the Model Provider rejects; or
fails because the Model Provider times out or cannot be reached.

6.5 Interrupted streams. If you disconnect during a streaming response, you are charged for the Tokens processed up to that point, because the Model Provider charges us for them. If the Model Provider does not report usage for an interrupted Request, we charge an estimate based on the input you sent and the output delivered before you disconnected.

6.6 Reservations. When a Request starts, we reserve an estimated amount of AI Credit to cover it. The reservation is released when the Request finishes, and you are charged the actual cost. A reservation can be larger than the final charge, so you may briefly see less available credit than you will end up spending.

6.7 Costs above the reservation. If a Request's actual cost exceeds its reservation, the full actual cost is still charged. This can take your available AI Credit below zero. While your available AI Credit is below zero, new Requests are refused. Your next AI Credit purchase covers the shortfall first.

6.8 Balance updates. Charges are recorded immediately against your available AI Credit, and posted to your billing records periodically. Your dashboard may briefly show a charge as pending before it is posted.

6.9 Price changes. We may change Model prices. A new price applies to Requests made after it takes effect, never to Requests already completed. We will publish price increases in your dashboard at least [7] days before they take effect.

7. Spending Controls

7.1 You can set a monthly budget and a lifetime budget on each API Key. A budget of zero means no budget. Monthly budgets reset on the first day of each month (UTC).

7.2 Budgets are checked before each Request. A Request that starts within a budget may finish slightly above it, and is charged in full. After that, further Requests on that key are refused until the budget resets or is raised.

7.3 Your AI Credit balance is always the final limit, whatever budgets you set.

7.4 Spending controls are tools to help you manage cost. You remain responsible for all charges made with your API Keys.

8. Refunds

8.1 AI Credit is a prepaid balance for usage that is processed and paid for in real time. AI Credit is not covered by the 30-day money-back guarantee in our Refund Policy.

8.2 Unused AI Credit is non-refundable, except:

where we have charged you in error, which we will correct by refund or by restoring AI Credit;
where we discontinue the AI Gateway, in which case we will refund any unused AI Credit; or
where applicable law requires a refund.

8.3 If you close your account, or we terminate it under Section 14.2, unused AI Credit is forfeited unless Section 8.2 applies.

8.4 Approved refunds are paid as described in our Refund Policy.

9. Acceptable Use

9.1 You must use the AI Gateway lawfully and in line with our Terms & Conditions.

9.2 You must also comply with the usage policies of the Model Providers whose Models you use, including OpenAI's Usage Policies (opens in a new tab). A use those policies prohibit is prohibited here.

9.3 You must not:

use the AI Gateway to create, distribute or facilitate illegal content, malware, fraud, phishing, spam, harassment, or content that exploits or harms children;
try to get around rate limits, spending limits, capacity controls, authentication, or any other restriction, including by using multiple accounts or keys to avoid a limit;
interfere with, overload or disrupt the AI Gateway or any Model Provider;
probe, scan or test the AI Gateway for vulnerabilities without our written permission;
send Input you do not have the right to send;
use Output to develop models that compete with a Model Provider, where that Model Provider's terms prohibit it; or
present Output as human-written where doing so would mislead people in a way that causes harm.

9.4 High-stakes decisions. Do not rely on Output alone to make decisions with legal, medical, financial, employment or similarly significant effects on people. Have a qualified person review it first.

9.5 We may investigate suspected violations of this Section.

10. Your Own Users

10.1 You may build applications on the AI Gateway and offer them to your own users ("End Users").

10.2 If you do, you are responsible for your End Users' use of the AI Gateway through your application as if it were your own. You must:

bind your End Users to terms at least as protective as Sections 9 and 12;
not give End Users your API Keys; and
suspend an End User promptly if we ask you to because of a violation, including at the request of a Model Provider.

10.3 You must have any rights, permissions and consents needed to send your End Users' data to the AI Gateway, and you must tell them that their Input is processed by third-party AI services.

11. Your Data

This Section supplements our Privacy Policy.

11.1 No training, no sale. We do not use your Content to train AI models, and we do not sell your Content.

11.2 Model Providers. To answer a Request, we send its Input to the relevant Model Provider. The Model Provider processes it under its own terms and data-handling policies, which we do not control. We do not send your account details or API Key to the Model Provider; we identify Requests to it only by an internal reference.

11.3 Sensitive data. Do not send passwords, payment card numbers, government identification numbers, health information, or other sensitive personal data unless you have a lawful basis and any consent required to do so.

11.4 Your rights. You can exercise your data rights by contacting support@n8nclouds.com. We will respond as described in our Privacy Policy.

12. Outputs

12.1 Ownership. As between you and us, you own your Input and, to the extent permitted by law and the Model Provider's terms, the Output generated for you.

12.2 Accuracy. Output is generated by AI and may be inaccurate, incomplete, offensive or out of date, even when it appears confident. Similar Inputs can produce different Outputs. You are responsible for evaluating Output before relying on it, and for how you use it.

12.3 No endorsement. Output does not reflect our views and is not professional advice.

13. Availability and Capacity

13.1 No uptime guarantee. The AI Gateway depends on third-party Model Providers. The Service Level Agreement in Section 7 of our Terms & Conditions does not apply to the AI Gateway, and AI Gateway downtime does not qualify for service credits.

13.2 Capacity. Model capacity is shared. When a Model is busy, a Request may wait briefly before it is processed, or be refused with a "busy" error that asks you to retry. Refused Requests are not charged.

13.3 Limits. Requests that exceed a Model's maximum context length, or that need more capacity than the Model can process at once, are refused.

13.4 Maintenance and outages. The AI Gateway may be unavailable for maintenance, or because of an outage at a Model Provider or one of our own providers. We may also temporarily disable the AI Gateway to protect the service or our customers.

14. Suspension and Termination

14.1 We may suspend an API Key, your AI Gateway access or your account, immediately and without prior notice, if:

we reasonably believe you or your End Users have violated these AI Terms or a Model Provider's policies;
a Model Provider asks us to;
an API Key appears to be compromised; or
it is necessary to protect the AI Gateway, our customers, or third parties.

14.2 We may terminate your AI Gateway access for serious or repeated violations.

14.3 While your access is suspended, your AI Credit is retained but cannot be used. Section 8 governs what happens to AI Credit on termination.

14.4 You can stop using the AI Gateway at any time by revoking your API Keys.

15. Limitation of Liability

15.1 The AI Gateway, Models and Output are provided "as is" and "as available", without warranties of any kind, including warranties of accuracy, fitness for a particular purpose, or uninterrupted availability.

15.2 We are not liable for:

Output, or any decision or action based on it;
the acts, omissions, outages or policies of Model Providers or other third parties;
usage of, or charges from, your API Keys by anyone who obtained them from you or your systems; or
indirect, incidental, special or consequential damages, or lost profits, data or business.

15.3 Our total liability arising from the AI Gateway is limited to the amount you paid for AI Credit in the 12 months before the event giving rise to the claim.

15.4 Nothing in these AI Terms limits liability that cannot be limited under applicable law.

16. Changes, Governing Law and Contact

16.1 Changes. We may update these AI Terms. We will notify you of material changes by email or in your dashboard before they take effect, except for changes needed for legal, security or Model Provider reasons, which may take effect immediately. Continuing to use the AI Gateway after a change takes effect means you accept it.

16.2 Governing law. These AI Terms are governed by the laws of Bangladesh, and disputes are resolved as described in Section 11 of our Terms & Conditions.

16.3 Contact.

Email (legal, privacy, support and compromised keys): support@n8nclouds.com
Phone: +880-1925-363333