Terms of Service

02 Platform Scope & Hugging Face Integration

Jutt AI Hub is an open-access artificial intelligence curation and security intelligence matrix. It provides real-time model discovery, technical parameter estimation, and execution guides by aggregating public data from Hugging Face.

• Third-Party Model Licensures: All artificial intelligence models, weights, fine-tunes, tokenizers, and datasets listed on Jutt AI Hub are the intellectual property of their respective creators and organizations. You are responsible for reviewing and complying with each model's individual license (e.g., Apache 2.0, MIT, Llama 3 Community License, Gemma Terms, DeepSeek License).
• No Direct Hosting of Weights: Jutt AI Hub does not independently host or re-distribute copyrighted model binary weights. When you download models via provided CLI snippets (such as ollama run or huggingface-cli), files are downloaded directly from the author's designated upstream repositories.
• API Availability: While we strive for continuous uptime, model metadata availability is reliant on public Hugging Face APIs and may experience occasional upstream latency or rate limits beyond our control.

03 Acceptable Use & Ethical AI Policy

As a cybersecurity entity, Jutt Cyber Tech™ strictly requires that Jutt AI Hub and any models discovered through it are used solely for lawful, defensive, academic, or legitimate commercial development purposes. You explicitly agree NOT to:

• Generate Malicious Payloads: Prompt or fine-tune models to generate weaponized exploit code, polymorphic malware, zero-day exploitation tooling, ransomware, or keyloggers targeted against unauthorized third parties.
• Unauthorized Attacks: Conduct automated vulnerability probing, denial-of-service (DoS/DDoS) campaigns, or credential stuffing operations against computer systems without explicit written consent.
• Disinformation & Non-Consensual Media: Generate non-consensual deepfake media, deceptive voice cloning for financial fraud, or automated mass disinformation campaigns.
• Harm or Harassment: Produce content promoting violent extremism, child sexual abuse material (CSAM), non-consensual surveillance, or targeted harassment.

04 Hardware Estimators & Technical Information Disclaimer

Jutt AI Hub provides technical estimations, including dynamic GPU VRAM calculators, context window scaling metrics, and quantized memory footprints (FP16, Q8_0, Q4_K_M).

• These calculations are provided for developer planning purposes based on theoretical neural parameter formulas and standard KV-cache allocations.
• Actual runtime memory consumption may vary depending on driver overhead, CUDA context allocation, activation caches, framework optimizations (e.g., FlashAttention-2, vLLM paging), and batch sizes.
• Jutt Cyber Tech™ is not liable for hardware instability, out-of-memory (OOM) crashes, thermal throttling, or hardware failures occurring on user devices.

05 Supply Chain Security & Deserialization Risks

You acknowledge that loading machine learning model weights from untrusted third-party public repositories presents inherent software supply chain risks.

• Legacy PyTorch pickle checkpoints (.bin, .pt, .pkl) can execute arbitrary system code upon unpickling.
• You agree that downloading, running, and testing third-party models is done entirely at your own risk.
• Jutt Cyber Tech™ strongly recommends verifying SHA-256 signatures and prioritizing Safetensors (.safetensors) and GGUF formats inside isolated virtualization or container sandboxes.

06 Disclaimer of Warranties & Limitation of Liability

TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW:

• "As-Is" Provision: Jutt AI Hub, its interface, and technical tools are provided on an "AS IS" and "AS AVAILABLE" basis, without warranties of any kind, whether express, implied, or statutory.
• No Generative AI Liability: Jutt Cyber Tech™ shall not be held liable for any direct, indirect, incidental, special, or consequential damages resulting from the use of, or inability to use, any AI models discovered on our hub, or from any hallucinations, code errors, or inaccuracies generated by third-party model weights.
• Indemnification: You agree to defend, indemnify, and hold harmless Jutt Cyber Tech™, its founders, engineers, and affiliates from any claims, damages, liabilities, and expenses arising from your violation of these Terms or misuse of artificial intelligence models.

07 Modifications, Governing Law & Contact

We reserve the right to modify or update these Terms at any time to reflect technological advancements, legal standards, or platform expansions. Continued use of Jutt AI Hub following updates constitutes acceptance of the revised Terms.