Privacy & licensing
The Free Agentic Publication Digester (FAPD) publishes a static website. This page states, for humans and for AI agents alike, exactly what that means for visitors — and under what terms the content may be reused.
What this site does not do
- No data collection. No cookies, no analytics, no trackers, no fingerprinting, no advertising.
- No tracking scripts. The digest pages are pure static HTML. The
live page (
/today.html) carries exactly one small inline script, and all it does is display each timestamp in your local time beside the published Eastern time. It makes no network request, sets no cookie, stores nothing, and reports nothing to us — the computation happens in your browser and stays there. With JavaScript disabled the page is identical minus the local-time hint. - No accounts, nothing transmitted. The live page's keyword filter is a plain HTML form whose state never leaves your browser: it is not submitted, not stored, and not readable by us. Nothing on these pages collects, transmits, or retains anything you type or click.
- No third-party requests. Pages load no external fonts, scripts, images, or embeds — your visit talks to this server and no one else.
The one thing that is recorded
Like effectively every hosted service, the web server infrastructure serving this site keeps standard access logs — the requesting IP address, requested URL, timestamp, response code, and User-Agent — used solely for security monitoring and abuse prevention (for example, rate-limiting and intrusion detection), retained briefly under routine log rotation, and shared with no one. TLS certificates are issued by Let's Encrypt; certificate issuance is public by design (Certificate Transparency logs), like every HTTPS site's.
The MCP service. Requests to https://fapd.info/mcp pass through
the same web server logs described above. The MCP service itself logs,
for each request, the time, the protocol method, the tool or resource
name, the response status and size, the processing time, and the
client's self-reported software name, and an opaque request identifier
the web server assigns to each request so the two logs can be matched
up. It does not log your IP address,
the arguments you sent, or the content of any request or response. The
web server in front of it keeps a separate access log for /mcp
(address, time, method, status, size, software name, the same request
identifier; never a request body) used for rate limiting and to block addresses that send many
rejected requests in a short time; those blocks expire on their own.
Logs rotate on the same schedule.
That is the entire data story. There is nothing else to disclose.
Licensing
- Digests, site pages, and explanatory documents: CC BY 4.0 (opens in a new tab) — reuse freely with credit to "FAPD — Free Agentic Publication Digester."
- The pipeline code: Apache-2.0, in the project repository.
- Quoted official government text is public domain (17 U.S.C. § 105) — it was never ours to license.
The attribution ask mirrors our own citation ethic: for factual claims, cite the underlying official source each digest item links to; cite FAPD for the aggregation. Machine-readable reuse notes ride in /llms.txt.
Our own outbound activity
The crawler that gathers this site's source material has its own transparency page: Our crawler.
Contact
The project contact address appears in every request our crawler makes and in the repository metadata.