FAQ

Honest answers to the questions a compliance officer actually asks.

If your question isn't here, email matt@regflow.example and we'll answer within a business day. Common ones get added back to this page.

§ Product basics

How is this different from the Federal Register Daily Digest?

The free Daily Digest sends a list of every document published that day. Useful for archival, useless for triage. regflow scores each document 1-5 for compliance impact, tags it with specific industry topics, and only delivers what crosses your watchlist threshold. A team using only the Daily Digest reads ~200 titles per day; a team using regflow Pro reads 5-15 alerts per day on average.

What is the materiality scale?

A 1-5 score calibrated for compliance triage. M1 is noise (typos, scheduling, ceremonial). M2 is minor (clarifications, format changes). M3 is notable (definition shifts, new exemptions). M4 is significant (new compliance obligations, deadline shifts, filing requirements). M5 is major (new substantive rules, major withdrawals, enforcement posture changes). The scale is consistent across documents and agencies — see /how-it-works for the full mandate.

How fresh are alerts?

The fetch + classify worker runs daily by default. Pro tier subscribers can request hourly runs. Federal Register publishes overnight; regflow Pro alerts typically land within 6 hours of the FR publication.

Can I subscribe via email instead of RSS?

Email digest delivery is shipping in Q3 2026. RSS feeds are live today and work in any reader (Feedly, Inoreader, NetNewsWire, Slack RSS bot). Pro subscribers will be migrated to email digests automatically when available.

Do you cover state law?

Today: Federal Register only (which includes some federal-state programs). State cannabis boards and state healthcare licensing boards are on the Q4 2026 roadmap. EU and UK are planned for 2027. See /sources for the full roadmap.

§ Data & accuracy

How accurate is the LLM classification?

We use GPT-5.5 with a strict materiality mandate. Spot-check verification on the live corpus shows >95% agreement with manual triage on M4-M5 (the alerts that matter most). M2-M3 boundaries are noisier but still meaningfully better than keyword filtering. We are building a held-out gold-labeled eval set for ongoing drift monitoring.

What if you misclassify a document?

Every alert links to the source Federal Register URL. Compliance officers verify in seconds. Misclassifications can be flagged via Pro-tier feedback and feed into a weekly prompt-calibration review.

Do you read the full document or just the abstract?

Currently the title + agency context + abstract for materiality scoring (which is sufficient for the M3+ decisions that drive alerts). Full-body fetch for high-materiality documents is on the Q3 roadmap — this lets the rationale cite specific operative provisions.

Where does the data come from?

Federal Register public API (federalregister.gov/developers/api/v1). Free, no scraping, no terms-of-service issues. Every alert URL points back to the original Federal Register document.

§ Technical

Is there an API?

Yes. /api/docs returns paginated classified documents as JSON. /api/watchlists creates and reads watchlists. /api/stats exposes live counts. RSS feeds are at /rss/<slug>. Pro tier includes JSON API access; Free tier is RSS-only.

Can I export to Excel or Sheets?

JSON API supports any tooling that consumes JSON. CSV export of matched documents is shipping with the email-digest milestone in Q3.

Webhook delivery for matches — when?

Waitlist signup webhooks are live today (Discord, Slack, generic). Per-match webhook delivery to Slack/Teams/Discord channels is shipping with the Pro tier in Q3.

Can I self-host?

The codebase is structured as a portable monorepo (Node 22, SQLite or Postgres, Hono web framework). Self-hosting is on the roadmap for Team-tier customers with sensitive sourcing. Email about Enterprise self-host arrangements.

What about historical data?

Pro tier includes 30-day history. Team tier includes full backfill from the date we started collecting (June 2026). Historical Federal Register data older than that is available via the public API at no cost; we can backfill on request.

§ Billing & support

Is there a free trial?

30 days on Pro, no credit card. After the trial you confirm by email and the founding $79/mo price is locked for 24 months.

How do you handle cancellations?

Cancel any time. Feeds remain live until the end of the paid period. No pro-rated refunds, but no surprise renewal charges either — Pro renews monthly only with explicit consent for the first 60 days.

What is your support SLA?

Free tier: best-effort, no SLA. Pro tier: 24-hour email response on business days. Team tier: 4-hour email response, plus a quarterly compliance briefing call.

Do you offer a non-profit or academic discount?

Yes. 50% off Pro for verified non-profits, academic researchers, and trade-association compliance committees. Email about discount eligibility.

Have a question this didn't cover?

Email goes straight to the founder. No CRM funnel, no auto-responder.

Contact