FSLY

Fastly, Inc.

One filing signal, as this feed published it.

FSLY Fastly, Inc. BEARISH conviction 1

2026-08-18: Lovett Scott R. (President, Go to Market) — sell-to-cover (tax withholding on a vest) — $427.2K of FSLY (open-market, Form 4, code S) [sold to satisfy tax withholding on an RSU vest per the leg's own footnote — timing and size set by the vest and the tax rate, not by the seller] [CORRECTION 2026-09-01: round 13 — this row published “2026-08-18: Lovett Scott R. (President, Go to Market) — discretionary sell — $427.2K of FSLY (open-market, Form 4, code S)”, and its own filing says the seller had no say in it. 0001517413-26-000246's footnote F1, cited by the disposal leg this row counts, reads: “Shares sold to satisfy tax obligations in connection with the vesting of previously granted Restricted Stock Units.” That is a sell-to-cover: the shares were sold because an award vested and a withholding rate said how many, so neither the DAY nor the SIZE was a decision, and `informed:1` — the strongest bearish qualifier this feed has — was asserted over money nobody chose to raise. THE SENTENCE PREDATES THE RULE THAT READS ITS OWN FOOTNOTE. The sell-to-cover family is not one pattern but a growing set of arms, each added because a live filer stated the same mechanism in words none of the existing ones could reach — five widenings since the class shipped, the most recent of them today. This row was published on 2026-08-21, the arm that classifies it arrived afterwards, and nothing re-read the filing in between. This row was already live at the last full-corpus sweep (2026-08-26), so its filing was re-read once — under a parser that could not yet see this — and passed. THAT IS THE FINDING THIS PASS EXISTS TO MEASURE AND IT IS NOT ABOUT THIS ROW: a widening only ever reaches the rows somebody goes back and re-reads, so every arm shipped since the last full-corpus pass left a shadow behind it, and until today nothing had counted them. NO DOLLAR FIGURE, DATE, SHARE COUNT, ACTOR OR DIRECTION CHANGES: $427.2K really did leave, the direction stays BEARISH by house policy because this is a mechanics relabel and not a re-reading of the trade, and the sentence above is what today's emitters produce from this filing re-read on 2026-09-01. AND IT WAS NOT THIS SELLER'S EVENT EITHER: 3 insiders of the same issuer sold at ONE IDENTICAL $28.60 on 2026-08-18 and every one of them lands in this same class — this row and signals id 17595, signals id 17605. One identical price across 3 separate filings is an issuer-run withholding run settled for several officers at once, not 3 people who each chose the same morning and were each filled at the same number. Original text in signal_events]

Trade date
2026-08-18
Filed
2026-08-20
Flagged
2026-08-21

SEC Form 4 on EDGAR permalink 0001517413-26-000246

Everything sibyl flagged on 2026-08-21

What these fields mean

Direction
Direction (BULLISH/BEARISH) classifies the event, not the price. It is not a forecast, and no scout's direction has been validated as predictive.
Conviction
Confidence is a heuristic conviction score, not a calibrated probability. On a single filing it is the scout's own 1-5 grade; on a row that rolls several filings together it is their SUM and has no ceiling, so it can read above 5. The stored number is published as it stands — never clamped, never hidden.
Dates
Dates are rendered exactly as filed. A sibyl row leads with the TRADE date the Form 4 reports; a pelosi row shows the member's trade date and the date the disclosure was filed, which are rarely the same day. "Flagged" is the day this feed first published the row.
Backfilled rows
A row marked "already old news when flagged" reached the feed after the fact. It is a record of a filing, not of a timely catch.
Sources
Every row links to the filing it was read from — an SEC Form 4 on EDGAR, or the disclosure document filed with the House Clerk or the Senate. A row whose source document this feed cannot link to is left off it.

Free to quote, cite, and reference with attribution to apibeary.com. Informational only — not investment advice. Signals describe observed filings, not recommendations.