> ## Documentation Index
> Fetch the complete documentation index at: https://vaquill.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Rights and licensing
> Straight answers on what you may do with the data: persist it, build embeddings and derived indexes, show it to your own users, redistribute it, and what you keep if you stop paying
Most legal-data vendors make you book a call to find out what you are allowed
to build. Here are the answers.
This page is a plain-language summary. The
[Terms of Service](https://www.vaquill.ai/terms), Section 21, is what actually
governs, and it controls if the two ever disagree.
## The short version
| Can I... | Answer |
| ---------------------------------------------------------- | -------------------------------------------------------------------- |
| Cache and persist responses in my own database? | **Yes**, including permanently, where that store powers your product |
| Build embeddings, vector indexes or other derived indexes? | **Yes**, for your own product |
| Show the legal text to my own end users? | **Yes**, with a credit unless your plan waives it |
| Keep what I cached from the API if I stop paying? | **Yes.** No deletion obligation |
| Keep a bulk delivery after the contract ends? | **No.** Delete the compilation within 30 days |
| Redistribute the statutory text itself? | **Yes**, without restriction. It is public domain |
| Resell the structured corpus as a dataset, feed or API? | **No**, not without a separate signed agreement |
| Ship it under my own brand? | **Under a white-label agreement**, yes |
| Train a model on it? | **Yes**, unless the model is a competing legal-data service |
## Why the text is unrestricted
The statutory, regulatory and constitutional text is US primary law, and US
primary law cannot be copyrighted. *Georgia v. Public.Resource.Org, Inc.*,
590 U.S. 255 (2020) settled it, holding that even annotations prepared under
contract for a legislature are uncopyrightable government edicts.
We claim no ownership of that text and nothing in our terms restricts your use
of it.
What is licensed to you is the **compilation**: the schema, the citations, the
hierarchy, the cross-references, the metadata and the assembled corpus. That is
the work product, and it is the only thing the restrictions below touch.
This is worth checking against any vendor whose corpus derives from a
commercial legal publisher's codification, because that layer can carry rights
the raw text does not. Ask where each jurisdiction was sourced from.
## What you keep if you leave
Responses you cached during a paid term are **yours to keep**. No deletion
obligation, and you may keep showing them to your customers. That includes
the citations, hierarchy and cross-references those responses carry, because
a cached answer is not usable without them.
Different, and deliberately so. On termination or expiry of a bulk
arrangement you **delete stored copies of our compilation within 30 days**.
There the corpus itself is the deliverable, so permanent retention would let
a single period buy it outright.
You still keep anything you had independently retrieved through the API, the
public-domain text, anything obtained under the Open US Law licences, and
Output already embedded in work product you delivered to your clients.
The public-domain text carries no restriction either way, and never did.
Retention is not a licence to redistribute. Whatever you keep, you may not
resell our compilation as a standalone data product, nor use it to build a
competing legal-data API or database.
## What you may not do
Without a separate signed agreement, you will not:
* systematically extract, bulk-download, scrape or accumulate the corpus for any
purpose other than powering your own end-user product
* resell, sublicense or redistribute the underlying legal data or our
compilation as a standalone dataset, database, feed or API
* offer the API, or a service substantially derived from it, under your own or a
third party's brand
* use the API or its Output to build, train or improve a product that competes
with our legal-research, statute-retrieval or legal-data API services
* reverse engineer the API or attempt to derive its models or training data
* exceed or circumvent your plan's rate limits or credit allowance, including by
spreading access across accounts
Each of those is available under a written agreement. They are commercial terms,
not prohibitions on principle.
## Attribution
Unless your plan waives it, display **"Legal data by Vaquill AI"**, linked to
vaquill.ai, somewhere a user can reasonably find it in any product where the
Output is shown to your end users. A footer, an about page or a data-sources
page is fine. One credit per product, not one per response. The console gives
you the exact snippet.
It is **not** required for internal or back-office use where nothing is shown
outside your organisation, or for evaluation.
Attribution can be waived commercially. Talk to us.
## The open dataset
Separately from the paid API, the whole corpus is published as
[Open US Law](/docs/api-guide/bulk-and-sql):
* **Public-domain text**: no restriction
* **Our compilation**: CC BY 4.0
* **The ingestion pipelines**: Apache-2.0
CC BY 4.0 already grants commercial use, redistribution, remixing and model
training, irrevocably, on every version we have published. If your plan is to
own the corpus outright, start there and you owe us nothing but a credit.
## Still unsure?
Describe what you want to build and we will tell you which of the above it lands
in, in writing, before you commit to anything.
[contact@vaquill.ai](mailto:contact@vaquill.ai)