Terms of Service

Last updated: January 2026

Key Points (Plain Language Summary)

  • We run the servers, not the AI. You're responsible for what you do with model outputs.
  • Community votes on the model each quarter. Voting = shared responsibility.
  • 3-month terms with re-consent. No auto-renewals, no zombie accounts.
  • Your prompts stay local. Inference happens on Vancouver Island, not third-party APIs.
  • Archive your data each cycle. We don't guarantee disaster recovery—back up locally.
  • Early adopter pricing. 2026 rates may not continue in 2027.

1. The Platform

The Factory ("the Platform") is a sovereign compute infrastructure operated by Humanjava Enterprises Inc. ("the Operator") based in British Columbia, Canada. The Platform provides shared access to GPU compute resources running open-source large language models (LLMs).

We are an infrastructure provider, not a content provider. The Platform hosts compute resources and open-source models selected by the community. We do not create, curate, or endorse the outputs generated by these models.

2. Eligibility

Age & Payment Authorization: By completing payment for membership, you represent that you are either (a) at least 18 years of age or the age of legal majority in your jurisdiction, or (b) have obtained authorization from a parent or legal guardian who accepts these Terms on your behalf.

The cardholder assumes responsibility for all activity conducted under the associated account. Any age-related liabilities are carried by the party providing payment authorization.

Export Controls & International Access: The Platform is operated from Canada and primarily serves users in Canada, the United States, and allied jurisdictions. Access from other regions may require secondary verification to ensure compliance with Canadian export control regulations and international sanctions. The Operator reserves the right to deny or revoke access to users in jurisdictions subject to trade restrictions or where providing AI compute services would violate applicable law.

3. Open Source & "As-Is" Disclaimer

All models hosted on the Platform are open-source software provided under their respective licenses (Apache 2.0, MIT, Llama Community License, etc.).

THE PLATFORM AND ALL MODELS ARE PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. The Operator expressly disclaims all warranties, whether express, implied, statutory, or otherwise, including but not limited to warranties of merchantability, fitness for a particular purpose, and non-infringement.

Model outputs may be:

You acknowledge that LLMs are probabilistic systems and accept full responsibility for how you use their outputs.

Infrastructure Disclaimer: The Platform is provided on a "best effort" basis. We do not guarantee any specific uptime, availability, or performance levels. Hardware failures, maintenance windows, and capacity constraints may affect service availability. The Operator is not liable for any losses arising from service interruptions.

Service Level Target: Our planned availability target is 90% uptime—not "five nines" (99.999%). This is a small-scale sovereign infrastructure, not a hyperscaler. We aim to provide reasonable data speeds between our servers and the first network hop. Beyond that point, any issues arising from DNS failures, DDoS attacks, ISP outages, or other internet routing problems are inherent risks of the public internet and are not managed or warranted by the Operator.

4. Community Governance & Shared Liability

The Platform operates on a community governance model. Each quarter, members vote to select the primary model that will be loaded on the GPU infrastructure.

By participating in the quarterly vote, you:

  • Acknowledge the known characteristics, limitations, and risks of the selected model
  • Accept that the community collectively chose this tool
  • Assume shared responsibility for the model's outputs during that quarter

If you do not vote:

  • You agree to abide by the majority consensus
  • You accept the associated liability protections
  • Your continued use of the Platform constitutes acceptance

What You Are Liable For: By voting (or accepting the majority decision), you acknowledge that you have been informed of the selected model's known capabilities and limitations. You accept responsibility for your own use of model outputs—including any decisions, actions, or content you create based on those outputs. The community's collective choice does not create joint liability between members; each user remains individually responsible for their own conduct.

Operator Veto: The Operator reserves the right to exclude any model from the ballot or override a vote result if a model poses unacceptable legal, safety, or operational risks. This includes models with known vulnerabilities, license conflicts, or outputs that could expose the Platform or its users to significant liability.

This governance model distributes informed decision-making to the user community. The Operator serves as an executor of the community's collective will within reasonable bounds, not as a service provider making unilateral content decisions.

5. Membership Terms

Fixed 3-Month Terms: All memberships operate on 90-day cycles. At the end of each term, you must actively renew and re-affirm these terms to continue access.

Renewal Window (Days 75–90): During the final 15 days of your term, you will be prompted to:

Why 3-Month Terms? By requiring periodic re-consent, we ensure you are always operating under current terms with full awareness of the current model's characteristics. There are no "zombie users" bound by stale agreements.

6. Data & Privacy

Your prompts never leave the island. All model inference happens on local GPU infrastructure on Vancouver Island, BC. Your conversations and generated outputs are processed and stored on servers owned and operated by Humanjava Enterprises Inc. located in British Columbia, Canada—we do not send your prompts or model responses to third-party APIs.

Third-Party Services: While model inference is local, the Platform may use third-party services for supporting functions such as authentication (login/SSO), website hosting, payment processing, and operational databases. These services handle account and transactional data under their respective privacy policies, but your conversation content and model-generated outputs remain on our local infrastructure.

What we store:

  • Account information (email, profile)
  • Conversation history (for your use only)
  • Voting records (for governance)

What we don't do:

  • Train on your conversations
  • Sell or share your data
  • Log prompts for monitoring

Data Portability: You may export your conversation history at any time in standard formats (JSON, Markdown).

Intellectual Property: You retain ownership of the prompts you submit and any outputs generated in response to your prompts. The Platform claims no intellectual property rights over your content. However, you acknowledge that similar outputs may be generated for other users submitting similar prompts, and no exclusivity is implied.

No Disaster Recovery Guarantee: The Platform does not guarantee backup, redundancy, or disaster recovery for user data. Hardware failures, natural disasters, power outages, or other force majeure events may result in permanent data loss. You are advised to archive your data locally at the end of each renewal cycle. There is no ownership or guarantee that data from a previous cycle will persist in a restorable way.

Force Majeure: The Operator shall not be liable for any failure to perform obligations due to circumstances beyond reasonable control, including but not limited to: natural disasters, acts of god, fire, flood, earthquake, pandemic, war, terrorism, civil unrest, power failures, internet outages, or government actions.

7. "Show Your Work" & Attribution

The Platform is designed as private infrastructure for your creative and professional work. Your prompts, outputs, and workflows remain your private data unless you choose otherwise.

Showcase Requests: From time to time, we may request permission to showcase work that:

Attribution Rights: If your publicly posted work clearly indicates Platform use—through distinctive model outputs, workflows, or your own disclosure—we reserve the right to acknowledge that our infrastructure was used in its creation. We will not claim ownership or creative credit, but we won't hide our role as the compute provider.

Discretion Priority: We understand that many users value privacy and competitive advantage. Any showcase requests will respect your timing and preferences. We're building infrastructure for serious work, not a content farm.

8. Disposition & Data Deletion

If you do not renew at the end of your 3-month term:

Day 91 — Disposition Mode:

6-Week Grace Period:

No Recovery: Once purged, your data cannot be recovered. This is by design—we practice data minimization to protect both your privacy and our liability exposure.

9. Acceptable Use

The Platform is a shared resource. You agree not to:

Escalated Access: If your usage patterns suggest you're treating the Platform as dedicated infrastructure (e.g., sustained high-volume API calls), we'll reach out. Heavy users may be asked to upgrade to the Saturator tier ($200/month) which provides dedicated compute windows.

Operator Discretion: The Operator reserves the right, at its sole discretion, to:

No refunds will be issued for terminations due to abuse. The Operator is not required to provide justification for refusing renewal, though we will generally communicate concerns before taking action.

10. Service Tiers & Liability

Early Adopter / Alpha Pricing: Current pricing reflects early adopter rates for the initial 10 founding members in 2026. After this cohort is filled, prices may increase at the Operator's discretion. Participation at 2026 rates does not guarantee continuation at the same price in 2027 or subsequent years. As prices are adjusted, these Terms will be updated to reflect current pricing and will maintain a record of historical prices within the preceding calendar year.

Regular Membership ($100/month)

  • Shared GPU access (queue-based)
  • Community-shared liability
  • Standard voting rights

Saturator Tier ($200/month)

  • Dedicated compute windows (booked hours)
  • Individual "leaseholder" liability
  • Priority during shared hours

The Saturator tier functions as a dedicated resource lease. During your booked hours, you are effectively renting the GPU—what you generate during that time is strictly your responsibility.

11. Limitation of Liability

TO THE MAXIMUM EXTENT PERMITTED BY LAW:

You acknowledge that the models may generate inaccurate, biased, or harmful content. You agree to hold the Operator harmless for any damages arising from model outputs, provided the Operator has acted in good faith as an infrastructure provider.

12. Indemnification

You agree to indemnify and hold harmless Humanjava Enterprises Inc., its officers, directors, employees, and agents from any claims, damages, losses, liabilities, and expenses (including legal fees) arising from:

13. Governing Law & Dispute Resolution

This agreement is governed by the laws of British Columbia, Canada.

Informal Resolution: Before initiating formal legal proceedings, you agree to first contact the Operator to attempt informal resolution. We commit to responding within 14 days and working in good faith to resolve disputes. Many issues can be addressed through direct communication.

Formal Proceedings: If informal resolution fails, any disputes shall be resolved in the courts of British Columbia, Canada.

For users in the European Union, this does not affect your statutory rights under GDPR. For users in other jurisdictions, local consumer protection laws may apply where they cannot be waived.

14. Changes to Terms

We may update these Terms from time to time. Because membership operates on 3-month cycles with mandatory re-consent, you will always have the opportunity to review current terms before renewing.

Material changes will be communicated at least 14 days before they take effect to active accounts within the platform's interface.

15. Contact

Questions about these Terms? Reach out to @vveerrgg on X or email legal inquiries to the Operator.

Humanjava Enterprises Inc.
Vancouver Island, British Columbia, Canada