LAST UPDATED · 31 JULY 2026

Terms of Service

These terms govern your use of LLM Forge ("the service"), operated by Strix Technology (Pty) Ltd. By creating an account or using the service you agree to them. If you are using the service on behalf of a company, you agree on that company's behalf.

1. Who we are

The service is operated by:

Strix Technology (Pty) Ltd
581 Opstal Street
Pretoria, 0184
South Africa
Reg. no. 2015/150595/07
VAT no. 4720285198
hello@llmforge.io

2. What the service is

LLM Forge is a control plane. It provisions GPU infrastructure inside your own cloud provider account (for example AWS or RunPod), deploys an open large-language model onto it, and gives you a private, key-authenticated endpoint. The infrastructure it creates belongs to you: it runs in your account, under your provider agreement, and your provider bills you for it directly.

There are therefore two separate bills: our subscription fee for the platform, and your cloud provider's charges for the compute you run. We do not resell compute, we do not mark it up, and we cannot refund it.

3. Accounts

4. Your cloud accounts

To use the service you connect a cloud provider account and authorize us to create, inspect, and destroy resources in it on your instruction. In doing so:

5. Subscriptions, trials, and payment

Payment for the platform is processed by Paddle, our merchant of record. Paddle handles checkout, invoicing, and applicable taxes such as VAT, and Paddle's buyer terms apply to the purchase itself. Displayed prices exclude VAT unless stated otherwise.

6. Estimates, budget caps, and timers

Cost figures shown in the service — hourly rates, monthly projections, spend-to-date — are estimates based on published on-demand prices from a curated catalog. They are not a quotation, and your provider's actual invoice is authoritative. Budget caps and self-destruct timers are safety mechanisms enforced against our estimates on a best-effort basis: they substantially reduce, but cannot eliminate, the risk of unexpected charges (for example if a provider API is unavailable, access was revoked, or prices change). You remain responsible for monitoring your cloud spend.

7. Models and third-party licenses

The models in our catalog are third-party open-weight models, each distributed under its own license (for example Apache 2.0 or a community license). The weights are downloaded onto your infrastructure. You are responsible for your compliance with the applicable model license and any usage policies it imposes, in light of how you use the model.

8. Acceptable use

You may not use the service to break the law, to infringe others' rights, to generate or distribute material that is illegal in your jurisdiction or ours, or to violate your cloud provider's acceptable-use policies. You may not probe, disrupt, or circumvent the platform's security or its plan limits, resell the service, or abuse free capacity (including any shared sandbox we host) through automation or rate-limit evasion. By design we cannot see the traffic between you and your deployed endpoint; acceptable use is enforced at the platform level.

9. Your content and data

Prompts and responses exchanged with your deployed endpoint travel directly between your client and your infrastructure. We do not receive, store, or train on them. What we do collect, and why, is described in our Privacy Policy, which forms part of these terms. You retain all rights to your content; we claim none.

10. Termination and deletion

11. Availability and changes to the service

We aim for high availability but the service is provided without an uptime guarantee at this stage. Deployments depend on your cloud provider's capacity, quotas, and APIs, which are outside our control — a provider having no available GPUs, or a quota of zero, is not a defect in the service. We may improve or change the service over time; we will not materially reduce what your plan includes mid-billing-period.

12. Disclaimers

The service is provided "as is" and "as available". To the maximum extent permitted by applicable law, we disclaim all implied warranties, including fitness for a particular purpose. Output produced by models you deploy is generated by third-party software running on your infrastructure; we do not warrant its accuracy and you are responsible for how you use it.

13. Liability

To the maximum extent permitted by applicable law: (a) neither party is liable for indirect, incidental, special, or consequential loss, or loss of profits, revenue, or data; (b) our total aggregate liability arising out of the service is limited to the platform fees you paid us in the 12 months before the event giving rise to the claim; and (c) we are not liable for charges billed to you by your cloud provider, subject always to section 6. Nothing in these terms excludes liability that cannot lawfully be excluded, including under the South African Consumer Protection Act where it applies.

14. Indemnity

You will indemnify us against third-party claims arising from your content, your use of deployed models in breach of their licenses, or your breach of section 8, except to the extent caused by our own breach of these terms.

15. Changes to these terms

We may update these terms. For material changes we will give you at least 14 days' notice by email or in the dashboard before they take effect; continuing to use the service after that constitutes acceptance. The "last updated" date above always reflects the current version.

16. Governing law

These terms are governed by the laws of the Republic of South Africa, and the courts of South Africa have jurisdiction, without prejudice to any mandatory consumer protections that apply in your country of residence.

17. Contact

Questions about these terms: hello@llmforge.io.