Data licensing / API docs
Licensee API
Base URL https://quantisnow.com/api/v1/licensing. JSON in UTF-8, times in UTC. Keys come from us with your contract; ask for a test key.
Authentication
Send the key as Authorization: Bearer qs_... (or x-api-key). Licensee keys are ordinary Quantisnow API keys (they start with qs_) with a data license attached. Keys in the query string are refused here, so they never land in proxy logs. Every response carries x-ratelimit-limit, x-ratelimit-remaining and, when your license has a monthly cap, x-quota-remaining.
GET /stories
Without since you get the newest stories, newest first. Pass the next_cursor you got back as since and you get everything after it, oldest first. Keep doing that and you never miss or repeat a story. Pages can come back shorter than the limit, even empty, while the cursor still moves: stories outside your license are skipped, not counted.
tickerone or more, comma separated, up to 50.sincea story id or next_cursor.limit1 to 100, default 50.typeany of sec_filing, insider_trade, fda, press_release, comma separated.
let cursor = loadCursor(); // null the first time
while (true) {
const url = new URL("https://quantisnow.com/api/v1/licensing/stories");
url.searchParams.set("ticker", "NVDA,SAP,ASML");
if (cursor) url.searchParams.set("since", cursor);
const res = await fetch(url, { headers: { Authorization: `Bearer ${process.env.QUANTISNOW_KEY}` } });
if (res.status === 429) { await sleep(Number(res.headers.get("retry-after")) * 1000); continue; }
const { data, next_cursor } = await res.json();
for (const story of data) upsert(story); // key on story.id
cursor = next_cursor; saveCursor(cursor);
if (data.length === 0) await sleep(15_000);
}GET /stories/:id
One story. Use it to pick up the AI write-up, which usually lands a few minutes after the story itself.
GET /activity
?ticker=AAPL&days=90 (up to 365). Daily counts of sec_filings, insider_trades, earnings, analyst_actions and other_news for one company, for activity charts and “busy today” badges.
GET /usage
Your key's calls and records per endpoint per UTC day for the last 31 days, plus the state of your webhook. These are the numbers we invoice from.
Story fields
- id
- Story id. Ids only grow, so the largest id you hold is a valid cursor.
- rights
- "owned" for SEC, FDA and our own items. "wire" only appears on licenses that include our write-ups of press releases.
- type
- sec_filing, insider_trade, fda, press_release
- category
- Form or topic, e.g. "SEC Form 8-K", "SEC Form 4", "Earnings".
- tickers
- Every company the story is about, with its name.
- published_at, discovered_at
- UTC. published_at is when the source stamped it (EDGAR acceptance for filings), discovered_at when we had it.
- headline
- Our AI headline when there is one, else the filing title. Never a press release headline.
- summary
- Our one-line AI summary, when it differs from the headline.
- analysis
- Our AI write-up: explanation, key_figures (label and value), positives, negatives. Null until it is written; refetch the story to pick it up.
- scores
- impact 1 to 5 (how likely to move the stock), sentiment 1 to 5 (very negative to very positive, for a shareholder).
- insider_trade
- Form 3, 4 and 5 stories: insider, role, trade_date, transaction codes, bought and sold dollars and shares, owned_after, trading_plan.
- price_reaction
- Add-on. Once the next session closes: session_date, timing (pre, intraday, post), move_1d_pct, spy_1d_pct, move_5d_pct. Null when your license does not include it.
- source
- name and url of the original record, e.g. the filing on sec.gov. url is null on wire stories.
- url
- The story on quantisnow.com, for a "read more" link.
Webhook push
Give us an HTTPS URL and we POST new stories to it, checked every minute, up to 200 a request. We hold a story until its AI summary is written, for at most 10 minutes, so most pushes arrive complete. Every request is signed to the Standard Webhooks spec with a secret we send you once.
Answer with any 2xx. Anything else, or no answer within 10 seconds, and we send the same batch again a minute later with the same webhook-id, so you can drop repeats. Stories wait in line while your endpoint is down; nothing is skipped. If it stays down for ten minutes we hear about it and get in touch.
import { Webhook } from "standardwebhooks";
const wh = new Webhook(process.env.QUANTISNOW_WEBHOOK_SECRET); // whsec_...
// Throws if the signature or timestamp is wrong.
const event = wh.verify(rawBody, {
"webhook-id": req.headers["webhook-id"],
"webhook-timestamp": req.headers["webhook-timestamp"],
"webhook-signature": req.headers["webhook-signature"],
});
// event.type === "stories.created"; event.data is an array of storiesErrors
Errors look like {"error": {"code": "rate_limited", "message": "..."}}.
- 400 bad_request
- A parameter is wrong. The message says which.
- 401 missing_key / invalid_key
- No key, one we do not know, a revoked key, or a key with no data license.
- 403 license_paused
- The license is paused. Write to us.
- 403 ticker_not_licensed
- Your license is limited to some tickers and none you asked for are in it.
- 404 not_found
- No such story in your license.
- 429 rate_limited
- Over your per-minute limit. Wait the seconds in Retry-After.
- 429 quota_exceeded
- Over your monthly quota. Retry-After counts down to the 1st (UTC).
What each license includes
Every story passes a rights check on our side before it is sent, by API or webhook. A standard license includes SEC EDGAR filings and the FDA records we track, which are public records, and everything we produce from them: titles we build from filing data, our AI headlines, summaries, write-ups and scores, and parsed insider trades. The price move after each story is an add-on.
Company press releases and other wire stories are not resold. Their headlines, text and links are never in a payload. Where our feed terms allow it, a license can add our own AI write-ups of press releases; those stories carry "rights": "wire", a headline we wrote, and no source link. Sources we add later stay out of every license until we have cleared them.
Display a “Data from Quantisnow” credit near the data. Analysis is generated by AI from the source record and is not investment advice.