About Aervik Labs

Aervik Labs builds narrow, reliable APIs for compliance, trade, and document intelligence teams — products that answer one specific question well instead of trying to be everything. We work on the parts of compliance and operations work that are clearly defined, repetitive, and genuinely tedious to do by hand: checking a certificate of insurance against a contract's coverage requirements, checking a vendor company against federal exclusion and sanctions lists, checking a shipment against a forced-labor or denied-party list, pulling structure out of a document. Each product is built around one real, recurring question, documented plainly, and priced so a small team can actually use it.

How we operate

Aervik Labs is AI-operated and human-directed. An AI operating agent researches, builds, documents, and runs our products day to day. A human principal is accountable for the business and makes the decisions that matter commercially and legally — what we charge, what we publish, what we spend, and what we don't ship.

We built the company this way deliberately, as a real test of how much of a software business an AI agent can run well. We disclose it plainly because you should always know who — or what — is behind a product you're paying for.

Where AI sits, and where it doesn't

The distinction that matters to you as a customer is between the data we return and the software around it.

Our reference data comes from official sources, not from a model. Federal exclusion lists, sanctions and denied-party lists, tariff schedules, environmental compliance records, and securities filings are pulled from the issuing government agency's own API, bulk file, or public dataset. We record the license or terms-of-use basis for every source we use. Our responses tell you which sources were actually checked, with a date for each. A match reported from a live source is a match in the government's data — never a model's guess — and where a source isn't live yet, the response labels it rather than letting a clean result imply a check that never ran.

AI does the engineering and the explaining. The agent writes and tests the code, writes the documentation, and monitors the services. In most of our products AI also does a specific, disclosed job — reading an uploaded document, detecting personal data in text, drafting a plain-English summary of a filing, or suggesting a 6-digit HS subheading with a confidence score. Where that happens, the product's documentation says so explicitly, and any number that needs to be reliable is computed by fixed rules rather than generated.

A human decides the consequential things. Pricing, what gets published, what gets pulled down, spending, and anything with a legal or ethical question attached goes to the human principal. That authority gets used: in September 2026 he reviewed the whole portfolio and permanently discontinued several products we had already built, because operating them responsibly would have meant becoming a different kind of company than this one. One product is finished but not yet open for signups; its page says why.

When we get something wrong

We will get things wrong. Government sources change formats without notice, datasets go stale, and our own code has bugs like anyone else's. What we commit to is straightforward:

• Tell us and we'll look at it. Email aerviklabs@gmail.com with what you saw and what you expected.
• If we're wrong, we fix it and say so. Corrections to data handling, coverage, or documented behavior are recorded on our changelog with the date, rather than quietly edited away.
• If we can't fix it, we disclose it. Where a data source has a gap we can't close — a state that publishes nothing automatable, a source that blocks programmatic access — we name the gap in the product's documentation and in the API response itself, instead of letting a clean result imply coverage we don't have.

What we won't do

We don't invent people. There are no fabricated founders, bylines, staff bios, testimonials, reviews, case studies, or customer counts anywhere on this site or in any of our listings. If we haven't got customers to point to yet, we say nothing rather than imply otherwise. Editorial content is published under the Aervik Labs name, never under an invented person's.

Legal entity and contact

Aervik Labs is a brand operated by EAC Ventures, LLC. EAC Ventures, LLC owns the business, its accounts, and its liabilities.

For anything at all — a correction, a security or privacy concern, a question about how a product works, or a question about how the company is run — email aerviklabs@gmail.com. See also our Security & Data Handling page, Privacy Policy, and Terms of Service.