SEC Marketing Rule · 206(4)-1

Publish with confidence, not counsel fees.

Paste a draft LinkedIn post, newsletter, or pitch deck. In about 30 seconds, Safe to Publish shows you every clause an SEC examiner might cite, the rule behind each flag, and a compliant rewrite that keeps your voice.

Start free trial →How it works
Corpus updated Apr 22, 2026·No card required·AES-256 at rest · TLS 1.3 in transit
~30s
Median review time
206(4)-1
Full rule corpus
7 yr
Max audit retention
0
Hallucinated citations
“

Safe to Publish is a review tool. It surfaces clauses an SEC examiner might cite. The firm owns the final compliance decision.We never claim a draft is "SEC-approved" — no tool can.

What we cite

Grounded in the rule text — not a model's recollection of it.

Every flag resolves to a chunk of the corpus below. If a flag cannot cite an authority, it is dropped before you see it.

Rule 206(4)-1 21 sections

  • (a)(1)Untrue statements of material fact
  • (a)(2)Unsubstantiated claims · reasonable basis
  • (a)(4)Benefits without fair treatment of risks
  • (b)(1)Testimonial & endorsement disclosures
  • (c)Third-party ratings
  • (d)(1)Gross vs. net-of-fees performance
  • (d)(2)Standardized 1/5/10-year periods
  • (d)(5)Extracted performance
  • (d)(6)Hypothetical / backtested performance

Staff FAQs 41 items

  • FAQ 8Predecessor performance — conditions
  • FAQ 12Testimonial compensation & conflict disclosure
  • FAQ 14Social-media channel scope for disclosures
  • FAQ 17Third-party rating date & period
  • FAQ 19Performance across advertising channels
  • FAQ 23Portable performance
  • FAQ 28Form ADV Part 2A cross-references

Risk Alerts 3 alerts

  • Dec '25Additional observations — testimonial, third-party rating, hyperlinked disclosure deficiencies
  • Sep '23Initial observations — Marketing Rule compliance gaps
  • Jun '23Examinations focused on the Marketing Rule
Why not just ask ChatGPT?

Because an examiner won't accept "the model said it was fine."

Generic legal AI

Confidently wrong is worse than silence.

  • Cites sections that don't exist, confidently. Anthropic studies peg hallucination rates at 15–27% for regulatory citations.
  • No awareness of the Dec 2025 risk alerts unless its training data is current.
  • Nothing deterministic. Ask twice, get two different answers, neither audit-logged.
  • Your drafts may train the next model. No DPA, no audit record of what you sent or what came back.
Safe to Publish

Cited corpus + deterministic checks + audit log.

  • Every flag cites a real rule chunk.Flags that can't cite an authority are dropped, not printed.
  • Deterministic patternscatch what the LLM misses: "guarantee", hyperlinked disclosures, missing 1/5/10-year windows.
  • Corpus refreshed within 7 days of new SEC risk alerts. Every review stamps the corpus version.
  • Signed audit log, retained 7 years. Hand it to the examiner unedited.
Three steps

Paste. Review. Ship.

Paste any draft

LinkedIn post, email newsletter, blog copy, pitch deck slide, or a full website page. Up to 10,000 characters on Solo, 20,000 on Practice, and 40,000 on Firm.

Review against the corpus

Claude matches the draft against every cited chunk of the rule + staff FAQs + risk alerts. Deterministic checks catch the textbook gotchas.

Decide and export

Apply a rewrite, dismiss with a note, or escalate to counsel. Every decision appends to your signed audit log.

Your next LinkedIn post, reviewed in about 30 seconds.

14-day free trial. No card required. Up to 10 reviews to decide if it earns its keep.

Start free trial →See pricing