How does the PLARV Argus Anchor Point work?

The Anchor Point Protocol identifies the last genuinely stable step in a training run, allowing researchers to restore optimizer states and weights to a known clean baseline instead of restarting from scratch.

What is the Forensic Social Hub?

The Forensic Social Hub is a collaborative workspace where investigators and researchers can audit training runs together, deep-diving into signal breaches and tracking resolutions across the fleet.

PLARV Argus | Active Monitoring Control Plane

GOVERNANCE_PROTOCOL_V3.94

CRITICAL NOTICE: THESE TERMS GOVERN ACCESS TO THE ARGUS CONTROL PLANE. THEY INCLUDE BINDING DISCLAIMERS REGARDING COMPUTE BUDGETS AND MODEL PERFORMANCE.

1. Acceptance of Terms

These Terms of Service ("Terms") constitute a legally binding agreement between Datawiser LLC ("Company", "we", "us", or "our") and the institutional entity or individual researcher ("User", "you", or "your") accessing the Argus platform and related services.

By initiating an SDK handshake, authenticating an API pulse, or accessing our web interface, you affirm that you have the legal authority to bind your institution to these Terms. If you do not agree to every provision of this document, you must immediately terminate all training sessions and cease use of our infrastructure.

2. Deterministic Responsibility

Argus provides high-fidelity monitoring and autonomous fail-safes (e.g., mode="AUTO"). However, the training of artificial intelligence models is an inherently stochastic and high-risk activity involving significant compute expenditures.

2.1 No Guarantee of Intervention

You acknowledge and agree that Argus is a monitoring instrument provided on a best-effort basis. We DO NOT warrant or guarantee that our engine will detect every instance of model collapse, loss divergence, or gradient explosion. Furthermore, we DO NOT guarantee that an autonomous intervention (e.g., triggering a SAVE_NOW checkpoint) will be successful in every training scenario.

2.2 Compute Budget Risk & Absolute Liability Disclaimer

THE FINAL AUTHORITY AND RESPONSIBILITY FOR MONITORING AWS, GCP, AZURE, OR ANY OTHER INFRASTRUCTURE COMPUTE SPEND RESTS SOLELY WITH THE USER. Datawiser LLC is NOT liable for any "runaway" compute costs, excessive billing from third-party cloud providers, or the exhaustion of research credits. This limitation of liability applies regardless of the cause, including but not limited to: bugs, latency, or outages within the Argus platform; failed autonomous interventions; or the compromise of your API keys and credentials.

2.3 "As Is" Architecture and Open Source Use

The Argus platform, its API, and any reference implementations or proof-of-concept (POC) scripts are provided strictly on an "AS IS" and "AS AVAILABLE" basis. To the maximum extent permitted by law, Datawiser LLC disclaims all warranties, explicit or implied, including warranties of merchantability, fitness for a particular purpose, and non-infringement. We are not responsible for any model corruption, hardware failure, or security vulnerabilities arising from your implementation.

3. License and Restrictions

Subject to your compliance with these Terms, we grant you a non-exclusive, non-transferable, revocable license to access the Argus platform for monitoring and intervention purposes.

You expressly agree NOT to:

Reverse-engineer, decompile, or attempt to derive the source code of the Data Acceptor logic or the Metadata Firewall protocols.
Utilize the Service to monitor training loops involving illegal datasets or activities prohibited by relevant international law.
Probe, scan, or test the vulnerability of our API ingress gates without explicit written authorization from our security core.
Redistribute forensic telemetry or dashboard insights to unauthorized third parties or competitive active monitoring services.

4. Limitation of Liability

TO THE MAXIMUM EXTENT PERMITTED UNDER APPLICABLE LAW, DATAWISER LLC SHALL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, OR CONSEQUENTIAL DAMAGES INCLUDING, BUT NOT LIMITED TO:

Loss of proprietary model weights or training datasets.
Model state corruption, weight degradation, or numerical instability.
Unintended training loop termination or interruption of compute.
Damage to institutional hardware or cloud compute instances.
Loss of funding, research progress, or model performance.
Unauthorized access to your telemetry buffers resulting from your failure to secure API keys or intervention secrets.

IN ALL EVENTS, OUR TOTAL AGGREGATE LIABILITY SHALL NOT EXCEED THE TOTAL FEES PAID BY YOU TO DATAWISER LLC IN THE TWELVE (12) MONTHS PRECEDING THE CLAIM.

5. Financial Protocol

Subscription fees are handled via our payment processor, Stripe. We operate on a strictly deterministic billing protocol: we offer a conditional refund within the first seven (7) days of activation, provided that you have consumed less than 10% of your plan's allotted "detect endpoint" usage (steps). If your usage exceeds 10%, or if seven days have passed, all fees are strictly non-refundable due to immediate infrastructure commitments.

Detailed cancellation procedures and refund conditions are governed by our Refund & Cancellation Policy, which is incorporated into these Terms by reference.

6. Data Stewardship & Intelligence

By utilizing the Argus platform, you grant Datawiser LLC a non-exclusive, perpetual, irrevocable, world-wide, royalty-free license to utilize aggregated and anonymized signal telemetry ("Anonymized Data") for the purposes of internal research, security hardening, and the training of our predictive intelligence models (specifically the Argus Cerebrum engine).

We affirm that this data usage is strictly limited to non-identifiable mathematical scalars and forensic histograms. We will never utilize your raw model weights, private datasets, or identifiable research parameters in this capacity.

7. Indemnification

You agree to defend, indemnify, and hold harmless Datawiser LLC, its affiliates, directors, and employees, from and against any claims, damages, obligations, losses, liabilities, costs, or debt (including but not limited to attorney's fees) arising from: (a) your use of and access to the Argus platform; (b) your violation of any term of these Terms; (c) third-party claims regarding unpaid or excessive cloud compute invoices arising from your monitored sessions; or (d) any claim that your training data or models violate a third party's intellectual property rights.

8. Dispute Resolution & Class Action Waiver

Binding Arbitration: Any dispute arising out of or relating to these Terms or the breach thereof shall be finally resolved by binding arbitration administered by JAMS in accordance with its Comprehensive Arbitration Rules and Procedures.

Small Claims Court Carveout: Notwithstanding the foregoing, either party may bring an individual action in small claims court, provided the matter falls within the court's jurisdiction.

CLASS ACTION WAIVER: YOU AND DATAWISER LLC AGREE THAT EACH MAY BRING CLAIMS AGAINST THE OTHER ONLY IN YOUR OR ITS INDIVIDUAL CAPACITY, AND NOT AS A PLAINTIFF OR CLASS MEMBER IN ANY PURPORTED CLASS OR REPRESENTATIVE PROCEEDING.

9. Governing Law & Force Majeure

These Terms shall be governed by and construed in accordance with the laws of the State of Utah, USA, without regard to its conflict of law principles. Datawiser LLC shall not be liable for any failure or delay in performance resulting from causes beyond its reasonable control, including, but not limited to, acts of God, natural disasters, internet backbone outages, cloud provider failures (e.g., AWS downtime), or governmental actions.