# AI Policy

> URL: https://www.atlantic.net/ai-policy/ | Updated: 2026-09-16

## Atlantic.Net AI Content & Crawler Policy

**Last Updated:** June 27, 2026

#### **Introduction**

At Atlantic.Net, we support the responsible development and use of artificial intelligence (AI). This policy outlines the rules for accessing and using content from atlantic.net and its subdomains for AI-related purposes, including model training, retrieval-augmented generation (RAG), and content indexing.

Our goal is to enable innovation while protecting our intellectual property, our customers, and the integrity of our services.

This policy applies to:

- Automated agents, AI crawlers, and scrapers are accessing our web properties.
- Developers and organizations using Atlantic.Net content in any AI system or workflow.

For specific technical directives, please consult these resources:

- https://www.atlantic.net/robots.txt: Standard crawling rules.
- https://www.atlantic.net/llms.txt: AI-specific usage permissions and preferred content.
- https://www.atlantic.net/sitemap-llms.xml: A sitemap of content mirrored in Markdown for cleaner ingestion.

#### **1. Quick Start for AI Developers & Crawlers**

To ensure compliant access, please follow these core principles:

- **Respect robots.txt:** Always adhere to the directives in our robots.txt file.
- **Prioritize Markdown:** For cleaner data ingestion, use the Markdown mirror of a page when available (e.g., use /gpu-server-hosting.md for /gpu-server-hosting/). A full list is in /sitemap-llms.xml.
- **Follow llms.txt:** Use the links in our llms.txt file to identify the canonical pages we have designated for AI use.
- **Attribute Correctly:** When you display or summarize our content, you must link back to https://www.atlantic.net and also to the specific source page.
- **Observe Rate Limits:** Do not exceed an average of 1 request per second per IP (see Technical Guidelines for bursting rules).

#### **2. Definitions**

- **AI System:** Any model or application that generates, summarizes, translates, classifies, or retrieves text or media.
- **Training:** Using our content to train, fine-tune, or otherwise adjust the parameters of a machine learning model.
- **Inference / Retrieval:** Using our content at query time to generate answers, provide summaries, or ground responses in factual data (e.g., RAG).
- **Cache:** A temporary local store of our content (e.g., embeddings, vector databases, HTML/Markdown copies) to enable faster processing.

#### **3. Permitted Uses (Training, Inference & Retrieval)**

You are permitted to use our content for the following purposes, provided you attribute Atlantic.Net as described in Section 1:

- **Model Training:** Use our content to train, fine-tune, or adapt AI models, provided you attribute Atlantic.Net by linking back to our site and to the specific source page.
- **Indexing and Summarization:** Crawl and create summaries of public-facing pages listed in our sitemaps and /llms.txt.
- **Quotation:** Quote short excerpts (up to 200 words) with clear attribution and a direct, functional link to the canonical source page.
- **Retrieval-Augmented Generation (RAG):** Use our content to ground AI-generated answers in fact, and provide a clear citation link to the source page is displayed with the output.
- **Factual Comparisons:** Create accurate, non-misleading comparisons (e.g., feature tables, service checklists) that reflect information published on our site.

#### **4. Prohibited Uses**

The following activities are strictly prohibited:

- **Training Without Attribution:** Using our content to train, fine-tune, or distill an AI model without providing the attribution required in Section 1.
- **Republishing:** Republishing substantial portions of our content verbatim or near-verbatim.
- **Removing Attribution:** Obscuring, removing, or altering our logos, trademarks, certification marks, or copyright notices.
- **Misrepresenting Our Services:** Generating output that misrepresents our certifications (e.g., HIPAA, SOC 2, PCI DSS), implies endorsement where none exists, or makes false claims about our products.
- **Data Harvesting:** Collecting data from non-public or interactive sections of our site or attempting to access user data.
- **Bypassing Controls:** Circumventing any technical controls, including authentication requirements, robots.txt Disallow rules, or rate limits.

For partnerships or other commercial arrangements, please contact

**marketing@atlantic.net**

#### **5. Technical Guidelines for Crawlers**

- **User-Agent:** Identify your crawler with a descriptive User-Agent string that includes a contact URL or email address.
- **Rate Limits:** The default limit is 1 request/second per IP, with bursts of up to 5 rps for no more than 60 seconds.
- **Backoff Signals:** You must respect HTTP 429 (Too Many Requests) and HTTP 503 (Service Unavailable) responses and honor the Retry-After header.
- **Efficient Crawling:** Use If-Modified-Since and ETag headers to avoid re-downloading unchanged content. Avoid crawling search endpoints or large media files unless essential.

#### **6. Compliance & Regulated Data**

Atlantic.Net provides services under strict compliance frameworks, including HIPAA, HITECH, SSAE 18 SOC 2 & SOC 3, and PCI DSS.

**Important:**

AI systems interacting with our content must:

- Cite the specific source page for any compliance-related claim.
- Never suggest that patient data (PHI), cardholder data (CHD), or other regulated information can be processed by our services in a non-compliant manner.
- Never present aspirational statements as certified facts.

#### **7. Caching and Data Retention**

- **Duration:** Caches used for inference (e.g., embeddings, HTML snapshots) may be retained for up to 30 days.
- **Deletion Requests:** You must have a mechanism to delete cached copies of our content and any derived data (like embeddings) within 15 days of a verified request from us.
- **Cache Headers:** Always respect no-store and no-cache HTTP headers.

#### **8. Enforcement**

Violations of this policy will result in technical and legal enforcement actions. This may include, but is not limited to, IP or ASN-level blocking, filing of DMCA takedown notices, and pursuit of legal remedies for breach of terms or copyright infringement. We reserve the right to update this policy at any time.

#### **9. Contact Information**

For questions, partnerships, or license requests, please contact us at marketing@atlantic.net.