> ## Documentation Index
> Fetch the complete documentation index at: https://docs.focusalpha.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Company News Articles
> The raw news feed by company ticker: individual articles with the model's direction, impact score and one-line reason for that name, plus a link to the original.
Company News Articles is the raw news feed for a company: the individual articles that mention or affect it, newest event first. Each observation is one article as it relates to one company, carrying the model's read for that name — direction, an impact score from 0 to 100, a one-line reason, the event type, the publisher, and a link to the original.
**Plan:** Fund · **Credits:** 1 per call
Use [Company News Events](/news/news-events) first unless you specifically want the separate write-ups.
## What You Can Use Company News Articles For
See every write-up behind an event rather than the strongest one. Compare how different outlets covered the same story. Find the article that carried a particular claim. Reach the events that reprice a position without naming it.
## The Article Text Is Not Here
Article text is **not stored and not returned**. You get the title, the publisher's own blurb where it may be republished, the one-line `why`, and the link. To read the article, follow the link.
`summary_withheld` set to true means a blurb exists that FocusAlpha may not pass on.
## The Feed Arrives in Two Strengths
This is the whole stream for a company, and it comes in two strengths.
Only items scoring 40 or above on the rule scorer are read by a model — roughly a fifth of the stream. The rest are returned too, with `read_by_model` false and `impact_score`, `direction`, `why` and `event_type` all null.
Those nulls exist because **nothing produced them**, never because a model read the article and had no view.
Do not describe an unread row as unimportant: it was never assessed. On the day after NVIDIA reported, 107 of the 166 articles carrying NVDA sat below the bar.
`score`, the rule pass on a 0 to 100 scale, is present on **every** row and is what to rank on when you want everything. `min_impact` filters to the read half.
## Ticker Is What the Model Wrote
`ticker` on a row is what the model wrote, **not a resolved identifier**. About 20% are not independently confirmed by the entity layer — `ticker_confirmed` is false — and a few name private companies with no symbol at all.
Never present an unconfirmed value as a tradable security.
## Direction and Impact Are This System's Reading
`direction` and `impact_score` are the model's judgment of what the article means for that name. They are not a price move, not a forecast and not a recommendation. `human_reviewed` marks rows a person has checked.
## Key Company News Article Fields
`channel` says what kind of source it is: `wire`, `sec_filing`, `exchange_filing`, `company_ir`, `earnings_call` or `primary_source`. `publisher` names the outlet. `event_type` is one of the 34 coarse families, matched with normalisation on both sides so a canonical name and a legacy spelling of the same family — `m&a`, `fda`, `litigation` — reach the same rows across the whole history.
`relation` is `direct` when the article names the company, or `knock_on` when the company is reached through a filed relationship, which makes the row an inference.
## Company News Articles Date Convention
`since` and `until` take ISO timestamps and filter the feed. Rows are ordered newest event first. As with the event feed, timestamps are on the article's own clock and their precision varies.
## Company News Articles Coverage
Coverage starts 2026-07-26. There is no history before it.
## Company News Articles Sources
Articles come from news wires, SEC filings, Asian exchange filings, company IR pages and earnings calls. Direction, impact score and the one-line reason are FocusAlpha's reading of the article, not the publisher's.
## Query Company News Articles
Use the company news endpoint for one company.
A `company_id`, ticker, CIK or ISIN.
`direct`, `knock_on`, or `both` to get each kind labelled. Ask for `knock_on` to answer what is repricing a position that nobody is writing about.
`bullish`, `bearish` or `neutral`.
One of the 34 coarse families.
Minimum impact score, 0 to 100. Direct rows only, and it restricts results to the model-read half of the feed.
Minimum knock-on magnitude, 0 to 100. Knock-on rows only, on a different scale.
ISO timestamp lower bound. `until` sets the ceiling.
Rows per page, up to 200. Page through with `cursor`.
## Related Datasets
See also [Company News Events](/news/news-events) for the deduplicated event view, [Macro and Policy News](/news/themes) for stories that name no company, and [Structured 8-K Events](/events/8k-events) for the filing facts behind a headline.
```bash cURL theme={null}
curl "https://api.focusalpha.ai/v1/companies/NVDA/news?min_impact=60&limit=1" \
-H "Authorization: Bearer $FOCUSALPHA_API_KEY"
```
```json Response (one row; reaches trimmed) theme={null}
{
"data": [
{
"news_id": "1538d77b885c4f76c41c3662bfdb901d980fe6c3",
"event_id": "NVDA:results:20260901",
"relation": "direct",
"ticker": "NVDA",
"direction": "bullish",
"impact_score": 70,
"ticker_confirmed": true,
"title": "Nvidia: Betting Against AI's Landlord Is A Widow-Maker Trade",
"event_type": "analyst",
"event_subtype": "analyst_opinion,analyst_price_target",
"subject": "NVIDIA Corporation",
"why": "An analyst is initiating coverage with a strong buy rating, highlighting revenue sharing, AI integration, and private equity backing as key drivers for NVDA.",
"human_reviewed": false,
"score": 57,
"read_by_model": true,
"link": "https://seekingalpha.com/article/4942244-nvidia-betting-against-ai-landlord-is-a-widow-maker-trade",
"channel": "wire",
"publisher": "seekingalpha.com",
"summary": null,
"summary_withheld": true,
"event_at": "2026-09-01T23:06:54.000Z",
"time_basis": "published",
"time_precision": "minute",
"is_scheduled": false,
"reaches": [
{
"via": "competition",
"why": "Nvidia's AI dominance and recurring revenue model intensify competition for AMD's AI accelerator products.",
"ticker": "AMD",
"direction": "bearish",
"magnitude": 60
},
{
"via": "product",
"why": "Private equity firms like BX benefit from the massive AI compute demand Nvidia is underwriting.",
"ticker": "BX",
"direction": "bullish",
"magnitude": 30
}
]
}
],
"next_cursor": "2026-09-01T23:06:54.000Z|1538d77b885c4f76c41c3662bfdb901d980fe6c3",
"symbols_searched": [
"NVDA",
"NVD.DE"
]
}
```