Coverage
What you get
| Item | Detail |
|---|---|
| Product code | X-NEWS |
| Coverage | Starter: headlines built from company press releases and SEC 8-K filings, with our own sentiment score. Pro: a licensed newswire with full text |
| Granularity | One record per article |
| History | 2 years of licensed news history on Pro |
| Delivery | REST; WebSocket stream on Business |
| Update speed | As items are published |
News is licensed content, and that drives the price everywhere. So we split it. Press releases and 8-K filings are public, and they carry the news that moves stocks most directly: earnings, guidance, deals, executive changes. That's Starter. A full commercial newswire costs real money to license, so it sits on Pro.
Schema
Schema
| Field | Type | Required | Description |
|---|---|---|---|
id |
string | yes | Our article ID |
timestamp |
string | yes | Published time, UTC |
updated_timestamp |
string | optional | Last update |
title |
string | yes | Headline |
summary |
string | optional | Teaser or description |
body |
string | optional | Full text, only where licensed |
url |
string | yes | Link to the original |
image_url |
string | optional | Lead image |
authors |
array of string | optional | Bylines |
source |
string | yes | Publisher name |
symbols |
array of string | yes | Tickers mentioned; may be empty for macro news |
topics |
array of string | optional | Tags or channels |
sentiment_score |
number | optional | -1 to 1 |
sentiment_label |
string | optional | bearish, neutral, bullish |
ticker_sentiment |
array | optional | Per ticker {symbol, sentiment_score, relevance_score} |
language |
string | optional | ISO code |
News records use symbols (a list) instead of a single symbol, because one article often mentions several companies.
Example record (illustrative):
{
"id": "nws_01JD2Q8V9K6X",
"timestamp": "2026-11-18T21:05:12Z",
"title": "NVIDIA reports quarterly revenue above estimates",
"summary": "Data center sales drove the beat; guidance for next quarter was raised.",
"url": "https://www.example.com/news/nvidia-q3-results",
"source": "Company press release",
"symbols": ["NVDA"],
"topics": ["earnings"],
"sentiment_score": 0.62,
"sentiment_label": "bullish",
"language": "en"
}
Sample data
Sample data
No free news sample yet. A sample file is available on request: send a few tickers and a date range and we'll return the matching filing headlines in this schema.
Use cases
Use cases
Earnings and 8-K alerts. Filter on your watchlist in symbols and on topics such as earnings. Push an alert when a new item lands, with the title and url.
Sentiment as a feature. Average sentiment_score per ticker per day and add it as a column next to your daily bars. Test whether a sharp negative day predicts the next session's return for your universe.
Event filters. Skip new trades in a stock for an hour after any item with sentiment_label other than neutral. A cheap way to avoid stepping in front of news you haven't read.
Research archives. On Pro, pull 2 years of licensed history for a sector and count items per day. News volume spikes often line up with volume and volatility spikes.
Code
Code examples
Example: latest items for two tickers.
curl "https://api.marketdata.co/v1/news?symbols=NVDA,AMD&limit=20" \
-H "Authorization: Bearer $MARKETDATA_API_KEY"
Example: Python, daily average sentiment per ticker.
import os
import requests
import pandas as pd
resp = requests.get(
"https://api.marketdata.co/v1/news",
params={"symbols": "NVDA", "start": "2026-10-01"},
headers={"Authorization": f"Bearer {os.environ['MARKETDATA_API_KEY']}"},
)
resp.raise_for_status()
news = pd.DataFrame(resp.json()["results"])
news["day"] = pd.to_datetime(news["timestamp"]).dt.date
print(news.groupby("day")["sentiment_score"].agg(["mean", "count"]).tail())
Pricing
Pricing
| Plan | News |
|---|---|
| Starter, $29/mo | Headlines from press releases and SEC 8-K filings, with our sentiment score |
| Pro, $79/mo | Adds a licensed newswire with full text and 2 years of history |
| Business, $249/mo | Adds the WebSocket stream |
News isn't sold as a dataset, because news licences usually forbid it. Among the providers we studied, news plans run from $0 to $1,250 a month.
Side by side
How we compare
23 of the 50 providers we studied sell news, and most of them resell the same few wires. We don't pretend to be a newsroom. Our starting point is what's public and fast, filings and press releases, tag it to the same tickers as our market data, and add a licensed wire on Pro with the price stated up front.
FAQ
Frequently asked questions
How is the sentiment score calculated?
sentiment_score comes from our own model, scored from -1 (bearish) to 1 (bullish), with sentiment_label as a simple bucket. Treat it as one input, and test it on your universe before you trade on it.
Do I get the full article text?
On Pro, for the licensed newswire, body holds the full text. Starter items link to the original press release or filing through url.
What time zone are timestamps in?
UTC, in ISO 8601. Many press releases go out just before the 9:30 New York open or just after the 16:00 close, so convert carefully when you line news up with bars.
Can I store and redistribute the articles?
You can store them for your own use under your plan. Redistributing licensed news is not allowed without a separate agreement, which is why it isn't sold as a dataset.
Is it real-time?
Items arrive as they are published. The push stream comes with Business; on Starter and Pro you poll the REST endpoint.
Which languages are covered?
English first. The language field is there for when we add other sources.
Keep reading
Related
- Asset hub: Cross-asset data
- Docs: API Reference: Fundamentals and SEC Filings
- Learn: Trading Around CPI, Jobs and FOMC Releases, How to Choose a Market Data Provider
- Related data: SEC filings API, Earnings calendar API