Scan any stock tickers and surface unusual options activity — contracts trading on abnormal volume versus open interest, big-premium prints, and net call/put sentiment. Turns raw option chains into a smart-money flow signal.
Ranking distortion on strikes with no open interest. A contract with
zero prior open interest used a flat "999" ratio, so a routine 500-lot on a
freshly listed strike ($50k) outranked a $14M print with ratio 8 by 40:1.
The score now log-damps both freshness and premium; freshness still leads,
money counts too. (unusual_score values change; the ordering is what
matters and it is now sane. vol_oi_ratio for zero-OI contracts is now
volume / 1, capped at 999, instead of a flat 999.)
README claimed "nothing unusual = never billed", which was no longer true
since the whole-chain sentiment fallback: a quiet run returns one
sentiment row (scope: "full_chain") per scanned ticker. The README now
says exactly that, and how to get a free status row instead
(include_sentiment: false).
Outage wait halved. The _SYMBOL URL variant only exists for a few
index symbols; it was retried with full backoff even after the plain URL
had failed on 5xx, doubling the wait (up to ~70 s per ticker) for nothing.
It is now tried only after a clean 404.
Status-row message no longer claims "no chain data" when there was chain
data but nothing unusual.
0.2.0 — 2026-08-28
Reliability fix — every run now returns a result.
Unusual options flow is by nature rare, so most scans found nothing that
cleared the thresholds and the Actor pushed nothing — which Apify
counts as a failed run even though the scan worked.
Now, whenever Include sentiment is on, every scanned ticker gets a
whole-chain call/put premium lean row (scope: full_chain) even when
no single contract is unusual — a useful read on its own.
If a run genuinely has no data at all (nothing unusual and no chain
data), it returns a single status row explaining what happened instead
of an empty dataset.
0.1.x — 2026-07 → 2026-08
CBOE option-chain scan for contracts trading on abnormal volume vs open
interest, weighted by dollar premium, with a per-ticker sentiment summary.
Per-ticker fetch/calc isolation and a row-by-row push_data fallback so
one malformed row or a storage hiccup can no longer fail a whole run.
