TERMS FOR THE PUBLISHED DATA ============================ This file governs the *data* in this project: the Parquet releases, the tables served by the API, and the values inside them. The *software* that collects and publishes that data is licensed separately under the Apache License 2.0; see LICENSE. This is deliberately not a single license identifier, because a single identifier would be a false statement about what we own. Read section 1 before reaching for a summary. 0. WHAT YOU ACTUALLY HAVE TO DO ------------------------------- For almost every reader: * The disclosure facts (who filed, who traded, what, when, in which statutory amount band) are in the public domain. We assert no copyright in them and we grant you no license in them, because none is needed. You do not need our permission and you owe us no attribution for them. See section 1. * Our own contribution (the schema, the field descriptions, the coverage and denominator tables, the normalization vocabularies, the selection and arrangement of a release) is released under CC0 1.0 Universal, to the extent we hold any rights in it at all. See section 3. * Two third-party upstreams may contribute rows to some tables and they require attribution. If you redistribute those tables, carry NOTICE and ATTRIBUTION.md with them. See section 4. * A federal statute restricts what *you* may do with a financial disclosure report, regardless of copyright and regardless of anything in this file. It is not ours to waive. See section 5. Read it. It is the only part of this file that can create a liability for you. * Citation is requested, not required: CITATION.cff. 1. THE FACTS ARE NOT OURS TO LICENSE (PUBLIC DOMAIN STATEMENT) -------------------------------------------------------------- This section is a statement, not a grant. The underlying documents are US government financial disclosure filings: House and Senate Periodic Transaction Reports published by the Clerk of the House and the Senate Office of Public Records, and Forms 3/4/5 and 13F published by the Securities and Exchange Commission. Two independent legal facts apply, and the second is the load-bearing one. (a) 17 U.S.C. 105 denies copyright protection to "any work of the United States Government," and *Georgia v. Public.Resource.Org, Inc.*, 140 S. Ct. 1498 (2020), confirms the bar reaches legislators. This argument is good but not airtight for the *document images*, because a disclosure form is completed by the filer, who may be a candidate or a private citizen rather than a federal officer or employee acting within official duties. (b) The facts themselves are not copyrightable at all. Under *Feist Publications, Inc. v. Rural Telephone Service Co.*, 499 U.S. 340 (1991), copyright never protects facts or data, only an author's original selection, coordination, and arrangement of them, and effort expended in collection ("sweat of the brow") earns no protection. Extracting the reported facts and discarding the filer's arrangement leaves nothing to infringe, however (a) resolves. Because of (b), we hold no copyright in the disclosure facts we publish. We therefore do not purport to license them to you. They are in the public domain in the United States. We make no representation about their status under the law of any other country, including jurisdictions that recognize a sui generis database right. 2. WHY THIS IS NOT SIMPLY "CC0" ------------------------------- A CC0 dedication is an act by a rights *holder*. Its own text is a waiver, and then a fallback license, of "Copyright and Related Rights" that the affirmer holds in the work. Applying CC0 to this dataset as a whole would therefore assert that we hold copyright in the disclosure facts and are generously giving it up. We do not hold that copyright. Nobody does. Asserting it in order to waive it would be a small false claim of ownership over the public record, and it is the kind of false claim a downstream user could reasonably rely on. It also does real harm in one direction: a reader who believes our waiver is what makes the data free will also believe that a different publisher, one who declines to waive, can lock the same facts up. That is exactly backwards, and *Feist* exists to say so. So the facts get a statement of public-domain status (section 1) and the thin layer we might genuinely own gets a real CC0 waiver (section 3). The combination is longer than one SPDX identifier and it is the honest shape. If a distribution channel forces exactly one identifier, use `other` and point at this file. Where `other` is not accepted, `CC0-1.0` is the closest available approximation and understates our carve-outs in section 4 rather than overstating our rights. 3. CC0 1.0 WAIVER OF OUR OWN CONTRIBUTION ----------------------------------------- To the extent, and only to the extent, that we hold any copyright, database right, or related right in the following, we dedicate it to the public domain under the Creative Commons CC0 1.0 Universal Public Domain Dedication (https://creativecommons.org/publicdomain/zero/1.0/legalcode), including the fallback license CC0 provides where a waiver is not legally effective: * the published schema, table layout, column names, enumerated value vocabularies, and field descriptions; * the normalization decisions and their outputs, including owner-code normalization and statutory amount-band canonicalization; * the coverage, denominator, and expectation tables, and the measured coverage figures in them; * the point-in-time structuring, including `published_at`, `clock_precision`, `revision`, and `supersedes`; * the selection, coordination, and arrangement of each dated release, and the manifests and checksums that describe it. This waiver is scoped on purpose. It is a true statement about material we plausibly authored. It is not a claim over the facts in section 1, and it is not a grant of anything covered by section 4. Timestamps we measured ourselves deserve a specific note, because they are the part of this dataset most likely to be mistaken for a proprietary asset. An HTTP `Last-Modified` value served by the House Clerk is the publisher's own measurement, not our authorship; we read it, we did not write it. Our contribution is the act of having asked in time, and that is effort, which *Feist* tells us earns nothing. Consider them public domain and take them. 4. WHAT THIS FILE DOES *NOT* COVER ---------------------------------- 4.1 Third-party rows that carry attribution obligations ....................................................... Some congressional trade tables may include rows derived from third-party parsers of the same public documents. Those parsers are permissively licensed, not public domain, and their licenses require that their copyright and permission notices travel with copies and substantial portions: * `kovagent/congresskit`, offered as MIT OR Apache-2.0. We take the MIT arm. Copyright (c) 2026 userFRM. * `kadoa-org/congress-trading-monitor`, MIT. Copyright (c) 2026 Kadoa. So the dataset is not uniformly public domain, and any statement that it is would be wrong. If you redistribute a table containing rows from either upstream, carry NOTICE and ATTRIBUTION.md with it. Full detail, including which projections are affected, is in ATTRIBUTION.md. 4.2 Sources used only to grade our work, and never republished .............................................................. Several projects are used as independent correctness references. We compare our extraction against theirs, we count where we disagree, and we publish the disagreement rate. **We do not publish their rows, and their licenses do not reach our output.** Do not infer from their appearance in our documentation that any of their terms apply to this dataset, and do not infer that we obtained a license from them, because for redistribution purposes we did not need one and in several cases could not have gotten one. Two are worth naming here because a reader could otherwise guess wrong in either direction: * `TattooedHead/house-stock-watcher-data` has a license field of literally null, meaning no license at all. Grader only. Its rows are refused at release time by an allowlist, not by convention. * `jeremiak/us-senate-financial-disclosure-data` is CC-BY-SA-4.0, which is a permissive-looking license and is nevertheless excluded. Its share-alike condition would propagate to the entire published dataset, which would in turn purport to impose share-alike on public-domain facts. Grader only. The full grader list, and the enforcement mechanism, are in ATTRIBUTION.md. 4.3 Document bytes we did not obtain from the publisher ....................................................... Where a document image was retrieved from a web archive rather than from the publishing agency, the archive's own access terms constrain that retrieval path. This does not change the copyright status of the document in section 1, and it is a reason not to make an archive the distribution source for republished bytes. See ATTRIBUTION.md. 5. A STATUTORY RESTRICTION WE CANNOT LICENSE AWAY ------------------------------------------------- This is the only part of this file that can create a liability for you, and it survives every waiver above. 5 U.S.C. 13107(c)(1) provides, in relevant part: It shall be unlawful for any person to obtain or use a report ... (A) for any unlawful purpose; (B) for any commercial purpose, other than by news and communications media for dissemination to the general public; (C) for determining or establishing the credit rating of any individual; or (D) for use, directly or indirectly, in the solicitation of money for any political, charitable, or other purpose. Paragraph (2) authorizes the Attorney General to bring a civil action and a court to assess a penalty of up to $10,000. Four things about this restriction: * It is statutory, not contractual, and it is channel-independent. It attaches to the report, so it is not avoided by sourcing the data from us instead of from the agency, or from a web archive instead of either. * It binds *you*, the person who obtains or uses the report. It is not a copyright, we do not hold it, and nothing in sections 1 through 4 waives it or could. * Our publication of this dataset was reviewed by counsel and cleared. That clearance is about our publication. It is not advice to you about your use, and we are not your counsel. * Consequently, if your intended use is commercial and is not news or communications media disseminating to the general public, get your own legal advice before you build on this. We are not in a position to tell you that you may. 6. NO WARRANTY -------------- THE DATA IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ACCURACY, COMPLETENESS, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT. IN NO EVENT SHALL THE PUBLISHER BE LIABLE FOR ANY CLAIM, DAMAGES, OR OTHER LIABILITY ARISING FROM, OUT OF, OR IN CONNECTION WITH THE DATA OR ITS USE. This is not boilerplate here. The dataset publishes measured coverage denominators precisely because it is incomplete, its coverage percentages are computed against a government index that is itself demonstrably a lower bound, and it republishes filer errors verbatim by policy rather than silently repairing them. Read the coverage table and the datasheet before relying on any figure. 7. QUESTIONS THIS FILE DOES NOT SETTLE -------------------------------------- Recorded here rather than papered over: * Non-US database rights. Section 1 speaks to US law. Whether the EU sui generis database right could attach to the compilation, and to whom, is unanalyzed. * Whether a filer could assert copyright in free-form text they wrote inside a disclosure form (a comment field, an attachment) as distinct from the reported facts. We publish source text in comment fields verbatim. No one has tested this. * Publicity and privacy interests, which are not copyright and are not addressed by any license. The government's own published record names dependent family members on some filings. We republish the public record as filed; see the datasheet, which flags this explicitly. SPDX-FileCopyrightText: 2026 Jack Lutz SPDX-License-Identifier: CC0-1.0 AND LicenseRef-Public-Domain-Statement