Terms of Use

Last updated: 2026-03-30 · Effective immediately for all new purchases.

Plain-language summary: LINUS-AI runs entirely on hardware you own or control. Your data never leaves your machine. Your licence tier sets a feature ceiling; your hardware RAM sets a practical model-size floor. Read §4 for the hardware table.

1. Parties and Scope

These Terms of Use ("Terms") form a binding agreement between you (an individual or legal entity, "Licensee") and LINUS-AI ("we", "us"), governing your use of all LINUS-AI software products, including:

LINUS-AI — private AI inference engine (core product)
LINUS-AI Medical (aiMED) — HIPAA-compliant medical AI server
LINUS-BizCore and related products

By downloading, installing, or activating any LINUS-AI product, you agree to these Terms. If you do not agree, do not use the software.

2. Licence Grant

Subject to payment of the applicable fee and compliance with these Terms, LINUS-AI grants you a limited, non-exclusive, non-transferable (except as stated in §2.3) licence to install and use the software in binary form only.

2.1 Tier-specific rights

Tier	Price	Seats	Key prefix	Feature ceiling
Community	Free forever	1	None required	Inference only · 5B model cap · 6 profiles · GUI
Professional	$249 perpetual + $99/yr updates	1	`LNAI-PRO*`	70B models (hw-limited) · Vault · Agentic · Mesh · Full API · 14 profiles
Team	$799 perpetual + $199/yr updates	5	`LNAI-TEAM`	All Professional + Federated learning · Blockchain audit · 5-seat mesh
Enterprise	$3,500–$6,000/yr (custom)	Unlimited	`LNAI-ENT*`	All Team + Air-gap · SSO/LDAP · Fine-tuning · Unlimited models & profiles
Medical Solo	$799 perpetual	1 facility	`LNMD-MED*`	HIPAA inference · Clinical roles · Audit PDF · Vault · 70B models (hw-limited)
Medical Enterprise	$5,000–$12,000/yr	Unlimited facilities	`LNMD-MEDE*`	All Medical Solo + Multi-site · Blockchain audit · SSO/LDAP · Air-gap · Unlimited models

2.2 Perpetual vs subscription

Perpetual licences (Community, Professional, Team, Medical Solo): buy once, own forever — updates are optional, not forced. You may use the version of the software you purchased indefinitely. Feature updates and new major versions require an active Updates Plan (billed annually). If your Updates Plan lapses, the software continues to function at the version you downloaded.

Annual subscription licences (Enterprise, Medical Enterprise): access continues for the paid subscription period. Upon non-renewal, the software reverts to Community tier behaviour. No refund is issued for the current paid period; access continues until period end.

2.3 Licence transfer

Professional, Team, Medical Solo, and higher tiers include one machine transfer per year. To transfer, deactivate via Settings → Licence → Deactivate machine, then activate on the new machine. Community licences are not transferable (no key required — simply install on any machine).

3. Restrictions

You may not:

Reverse-engineer, decompile, or disassemble the software or its licence enforcement mechanisms.
Circumvent, patch, or disable feature gates or licence key validation.
Redistribute, sublicense, or sell the software or its binaries without written permission.
Use a Medical (LNMD-*) key in the core LINUS-AI binary, or a core (LNAI-*) key in LINUS-AI Medical — keys are product-specific and will be rejected.
Use the Community tier for commercial purposes if your organisation's annual revenue exceeds USD $100,000. Above that threshold, a paid tier is required.
Remove, obscure, or alter any copyright or licence notices embedded in the software.

4. Hardware tier limits and model size caps

Important: Your licence tier is a feature ceiling. Your hardware RAM is an independent practical floor. Even an Enterprise licence cannot make a 70B model run on an 8 GB laptop. The software enforces both limits simultaneously to prevent out-of-memory crashes.

Model size caps are based on conservative estimates for 4-bit quantised (Q4_K_M) models, which represent the typical deployment for local inference. If you use higher-precision models (FP16, Q8), effective limits are lower. GPU VRAM supplements RAM for inference; the software detects available VRAM and adjusts accordingly.

Hardware class	RAM range	Effective model cap (Q4)	Typical devices
Entry Consumer	≤8 GB	~3B params	Budget laptops, older MacBook Air, Raspberry Pi 5 (8 GB)
Mid Consumer	8–16 GB	~8B params	MacBook Air M1/M2, mid-range Windows laptops
High Consumer	16–32 GB	~14B params	MacBook Pro M2/M3 14", high-end Windows laptops
Professional	32–64 GB	~34B params	MacBook Pro M-series Pro/Max, workstation laptops
Workstation	>64 GB	Up to licence ceiling	Mac Pro, Mac Studio Ultra, multi-GPU workstations

4.1 Combined enforcement example

A Professional licence (licence ceiling: 70B) on a Mid Consumer machine (16 GB RAM, hardware cap: ~8B) will cap the usable model size at ~8B. Upgrading to a Professional machine (32–64 GB) raises the effective cap to ~34B. The software communicates this via the Hardware Passport panel on first run.

4.2 Mesh / distributed inference

Team and higher tiers include mesh networking. When models are distributed across multiple mesh nodes, the combined RAM of participating nodes determines the effective model cap — not any single node's RAM. This is the primary intended way to run 70B+ models on consumer hardware clusters.

5. LINUS-AI Medical — additional terms

HIPAA Notice: LINUS-AI Medical is a software tool; it is not a Covered Entity. Deploying organisations bear full responsibility for their own HIPAA compliance, BAA obligations, and PHI handling policies. LINUS-AI does not store, transmit, or have access to any patient data.

5.1 Scope of medical licence

Medical Solo licences are issued per facility. A "facility" means a single physical or logical deployment site (one server, one network segment). Deploying the software at multiple sites or across networks requires a Medical Enterprise licence.

5.2 Privacy officer configuration

The deploying organisation must designate a Privacy Officer and configure their identity in config/server.toml. LINUS-AI does not and cannot act as privacy officer; this responsibility rests entirely with the deploying healthcare organisation.

5.3 Audit trails

LINUS-AI Medical generates an append-only JSONL audit log of all inference events. This log is stored locally on the deploying server and is never transmitted to LINUS-AI. Medical Enterprise customers additionally receive a blockchain-backed audit ledger and exportable signed PDF compliance reports.

5.4 Clinical use disclaimer

LINUS-AI Medical is a decision-support tool only. Outputs must be reviewed by qualified clinical personnel before informing patient care. LINUS-AI makes no warranty as to the clinical accuracy of model outputs and accepts no liability for clinical decisions made in reliance on the software.

6. Trial terms

All LINUS-AI products include a 30-day free trial from the date of first installation on a machine. During the trial:

Core product: Professional-tier features are available (including Vault, Agentic, Mesh, API).
LINUS-AI Medical: Medical Solo-tier features are available (HIPAA audit, clinical roles, 70B hw-limited).

After 30 days without a paid key:

Core product: reverts to Community tier (5B cap, 6 profiles, no Vault/Mesh/API).
LINUS-AI Medical: all /api/v1/* clinical inference endpoints are locked (HTTP 402). Admin endpoints remain accessible for key activation.

Trial periods are per-machine (tracked by install timestamp). Re-installation on the same machine does not reset the trial.

7. Payment, refunds, and renewals

Payments are processed via PayPal (including credit/debit cards). Enterprise and Medical Enterprise customers may request invoice or bank transfer by contacting support@linus-ai.com.

Perpetual licences: 30-day money-back guarantee, no questions asked. Email support with your order reference.
Annual subscriptions: cancel anytime; no refund for the current paid period. Access continues to end of period. EU/UK consumers retain statutory withdrawal rights.
Updates Plans: annual renewal. Lapsed Updates Plans do not disable the software — only new feature updates are blocked.

License delivery and cancellation: Once a purchase is completed and your license key has been dispatched to the email address you provided, the order is considered fulfilled. You may cancel future renewal plans (Updates Plans or annual subscriptions) at any time, but no partial refunds are issued for any portion of the current paid period already in progress. The 30-day money-back guarantee for perpetual licences applies from the date of purchase, not the date of first use.

8. Warranty disclaimer and limitation of liability

THE SOFTWARE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, LINUS-AI SHALL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES ARISING FROM USE OF THE SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

LINUS-AI's maximum aggregate liability to you under these Terms is limited to the amount you paid for the software in the 12 months preceding the claim.

9. Privacy, telemetry, and internet-capable features

LINUS-AI software operates entirely on hardware you own or control in its default configuration. No data, prompts, model outputs, audit logs, or usage metrics are transmitted to LINUS-AI or any third party by default. The only network activity initiated by the software in default mode is:

Licence key activation (POST to license.linus-ai.com) — sends only the key and a machine identifier.
Model downloads from sources you explicitly configure.
Mesh peer discovery on your local network via mDNS (LAN only; never WAN).

Important — Internet-capable features: LINUS-AI is powerful, configurable software. Depending on your plan tier and settings, it may be configured to:

Perform web searches as part of agentic or RAG workflows (tool-calling feature).
Connect to internet-based AI token services or external APIs that you configure (e.g., OpenAI-compatible endpoints, third-party model providers).
Participate in the LINUS-AI blockchain audit network or peer mesh over WAN when explicitly configured.
Route inference through mesh nodes on external networks when mesh WAN mode is enabled.

When any such internet-capable feature is enabled by the user, data relevant to that feature may leave your hardware and be subject to the privacy policies of the third-party service. You are responsible for ensuring any enabled internet features comply with your applicable privacy obligations (including HIPAA for aiMED users). By activating such features, you acknowledge and accept this.

The LINUS-AI website uses standard web analytics (no personal data sold or shared). See our Privacy Policy for details.

10. Governing law and disputes

These Terms are governed by the laws of the State of Delaware, USA, without regard to conflict of law provisions. Disputes shall be resolved by binding arbitration under JAMS rules. Class action waiver applies to the maximum extent permitted by law. EU/UK consumers may retain the right to bring claims before local courts under applicable consumer protection law.

11. Changes to these Terms

LINUS-AI may update these Terms. Material changes will be announced via the website and, where possible, via in-product notification. Continued use of the software after the effective date constitutes acceptance of the revised Terms.

12. Contact

Questions about these Terms: support@linus-ai.com