Using Publiusly to Verify Civic Claims
How to use Publiusly's articles, feed, and receipts panels to check a claim about Cleveland city government against the primary source.
Updated
Every Publiusly article is written from a primary-source record and ships with a panel linking straight back to it. The point of this guide is not to convince you to trust the article — it is to show you how to not have to. Treat every Publiusly article as a starting point for finding the record, not as the final word on what it says. Here is exactly what Publiusly is, what it is not, and how to use it to get from a headline to the underlying record in a couple of clicks.
What a Publiusly article actually is
Publiusly ingests Cleveland City Council's Legistar record — currently the only source it pulls from — through a nightly automated process that normalizes new and updated legislative items and stores them. That record is the raw material for everything else on the site: the feed, the legislation and meeting pages, and the AI-written articles.
The articles themselves are written by an AI system, under the byline "Publius" — an explicitly stated AI persona, not a real reporter. Publius drafts each article from the underlying Legistar record, including text extracted from attached documents where possible, and every article carries an accuracy disclaimer stating plainly that the analysis is algorithmically generated, may contain errors or omissions, and that Publiusly does not guarantee its accuracy — always verify against the original source records. That disclaimer is not boilerplate to skip past; it is the operating instructions for how to read the site.
Some articles also draw on Brave Search results as outside enrichment — added context from the open web about a named program, organization, or topic mentioned in the record. That enrichment can be useful for background, but it is a secondary layer on top of the primary-source record, not a replacement for it, and it is worth mentally separating "what the record says" from "what outside search turned up" as you read.
What it is not
Publiusly is not a substitute for reading the record yourself, and it does not claim to be. It is not a wire service with a human reporter who called the department for comment. It is not a legal or investigative authority, and nothing in an article should be treated as a final determination of fact about a person, an organization, or a legal question. Coverage is currently limited to one governmental body — Cleveland City Council — so an absence of an article about something is not evidence that nothing happened; it may simply mean the record has not been ingested, or the topic falls outside council's business entirely.
The honest way to use Publiusly is as an orientation and discovery tool: it helps you find the record you didn't know existed and gives you a plain-language starting point for understanding it. The verification step — actually confirming a specific claim — still belongs to you, and the site is built to make that step fast rather than to make it unnecessary.
The receipts panel: what it contains
Every Publiusly article carries a panel labeled "Check the receipts." That panel is the bridge between the article and the primary source, and it contains, at minimum: the file's title, its file number, its matter type (Ordinance, Emergency Ordinance, Resolution, or Communication), its current status, the committee it was referred to, an internal link to Publiusly's own legislation detail page for that file, a "View on Legistar" link that goes out to the official public record, and links to the source documents the article was written from.
That external Legistar link is worth trusting specifically because of a technical detail covered in the companion guide on Legistar file numbers: Cleveland's Legistar setup runs two backends with disagreeing internal IDs, and a naively constructed link can land on the wrong record entirely. Publiusly resolves that mismatch before publishing the receipts link, so clicking through takes you to the correct file rather than a coincidentally similar one.
How to actually check a claim, step by step
The workflow is short. Say an article makes a specific claim — a dollar figure, a program name, a neighborhood, a department. Here is how to run it down:
- Open the "Check the receipts" panel on the article and note the file number and matter type.
- Read the source document links in the panel, if any are listed — this is the attachment text the article was actually generated from, and the fastest way to see whether a specific claim traces back to specific language in the record.
- Click "View on Legistar" to open the official public record for that file number and compare the title, status, and committee against what the article said.
- If the claim still is not resolved by the record itself — for instance, a question about current departmental practice that goes beyond what is in this specific file — treat that as a question for the relevant city office directly, not something Legistar or Publiusly can settle on its own.
A worked example
Publiusly published an article headlined "New Home Repair Initiative Coming to Slavic Village," at this Publiusly article, built from Emergency Ordinance 882-2026 — the file authorizing the Department of Aging to contract with Community Housing Solutions on a home repair program in Slavic Village, funded through Closed Ward 12 Casino Revenue Funds. (The full clause-by-clause breakdown of that ordinance's title is in the companion guide on reading a Cleveland ordinance.) That article's receipts panel gives you the file number, the matter type, the status, the committee, and a corrected link to the Legistar record — so if you wanted to confirm, say, exactly which fund is paying for the program, you could go straight to the source document listed in the panel rather than taking the article's word for it.
If something looks wrong
Publiusly articles carry a feedback control — helpful or not helpful — and sustained negative feedback on an article automatically re-queues it for a rewrite rather than leaving a flagged error to sit indefinitely. If you spot something that looks inaccurate or incomplete after checking it against the record, using that control is the direct way to flag it. You can also reach Publiusly through the contact page for anything the feedback control does not cover.
If the receipts panel and the article disagree, trust the primary record — then tell us so we can fix the article.
This is also a good moment for a broader habit: treat any AI-generated summary of a government record — on this site or anywhere else — the way you would treat a secondhand account of a meeting you did not attend. Useful for finding out what happened and where to look, not a substitute for the transcript.
Where to go from here
Beyond individual articles, Publiusly exposes the underlying record directly, without AI summarization in between, if you would rather start there: the feed shows newly ingested items as they come in, the legislation and meetings pages under Cleveland City Council let you browse and search the record itself, the sources page shows where enrichment context comes from, and the visualizations page shows aggregate patterns across the legislative record. The how it works page covers the full pipeline — ingestion, document extraction, AI synthesis, and the receipts system — in more technical detail than this guide.