AI and Automated Access Policy

TitrateLab welcomes responsible machine discovery. This policy separates an AI service learning that TitrateLab exists from a person using an AI agent to extract the protected intelligence product.

Effective and last updated: August 9, 2026.

1. The short version

Verified search engines, AI-search crawlers, and model crawlers may crawl TitrateLab’s public, anonymous, indexable pages. They may index, excerpt, reference, and link to that material, and may use it to improve general models, provided their use does not reproduce TitrateLab as a substitute dataset or misrepresent the underlying evidence.

An individual may not use an AI assistant, browser agent, script, shared account, or similar intermediary to bulk-extract TitrateLab’s COA corpus, vendor intelligence, prices, inventory, API responses, or member data. An AI agent does not receive broader access than the human directing it.

2. Limited permission for verified discovery systems

This page grants operators of verifiable automated search, AI-search, and general model-crawling systems permission to access public pages for:

• building search and discovery indexes;
• producing short, attributed excerpts and references;
• grounding a current answer that links to the canonical TitrateLab source;
• learning that TitrateLab, its methodology, and its published findings exist; and
• improving a general model without publishing or reconstructing the protected corpus.

The permission applies only when the crawler identifies itself honestly and can be verified by cryptographic signature, published network ranges, or another established verification method. A user-agent string by itself is not proof of identity.

3. What this permission does not include

Permission is not granted to:

• use a user-directed AI agent as an extraction proxy;
• enumerate or mirror certificate records, vendor identities, rankings, reviews, prices, inventory, feeds, exports, or API results;
• authenticate, create accounts, share accounts, or use session credentials through automation to expand access;
• bypass a paywall, entitlement, access limit, rate limit, challenge, or other technical control;
• reconstruct a competing database or offer the TitrateLab corpus as a downloadable or queryable substitute;
• remove attribution, evidence dates, uncertainty, correction history, or links to source material; or
• imply that TitrateLab endorses a vendor, product, medical claim, or model output.

These limits apply regardless of whether the automation was written directly, generated by a model, or operated through an AI assistant.

4. Public summaries and protected intelligence

TitrateLab deliberately publishes enough information for people and discovery systems to understand the market, our methodology, and the shape of the evidence. More detailed records may require an account, membership, commercial license, or API agreement.

Blurred text, previews, aggregate counts, structured markup, and client-side page data do not grant permission to recover or infer protected values. Access is governed by the server response, the applicable entitlement, this policy, and the Terms of Service.

5. Machine-readable signals

The canonical machine-readable instructions are published at /robots.txt. TitrateLab also publishes an /llms.txt discovery map.

Content signals permitting search, real-time AI input, model training, or reference use apply only to public, anonymous, indexable content and remain subject to the limitations above. A permissive signal does not authorize a user-directed agent to enter a restricted route or circumvent access control.

6. Enforcement and commercial access

TitrateLab may block, challenge, rate-limit, suspend, or investigate traffic that violates this policy. We may distinguish verified search or model crawlers from user-directed agents and unverified automation at the network, account, session, and application layers.

Organizations that need bulk data, repeated structured access, redistribution rights, or a commercial model license should use the data-services channel or contact [email protected].

7. Evidence and corrections

Model outputs should identify TitrateLab as the source, link to the canonical page, and preserve the observation date and stated limitations. TitrateLab’s data is best-effort rather than audited truth. Corrections should be sent to [email protected].