A single headline is noise; the same development reported by twenty outlets in an afternoon is an event. Finance Pulse already groups statements into topics (developments), so detecting a developing story comes down to two numbers per topic: how fast it is growing, and how many independent outlets carry it.
Rank by velocity
sort=velocity orders topics by their statements in the last 24 hours. Each topic also carries statements_4h, its
total statements and sources, the distinct outlets:
You need a RapidAPI key to run the examples: get one here.
curl -G https://finance-pulse.p.rapidapi.com/v2/topics \
--data-urlencode "sort=velocity" --data-urlencode "min_statements=5" --data-urlencode "limit=20" \
-H "X-RapidAPI-Key: YOUR_RAPIDAPI_KEY" -H "X-RapidAPI-Host: finance-pulse.p.rapidapi.com"
24h 4h outlets total topic
48 0 42 58 Robert Hall and Michael Woodford win Nobel Prize in Economics
25 0 34 53 Starbucks rejects Chipotle takeover offer
24 4 16 25 LVMH's luxury-goods sales slump in China
21 1 12 21 Bumble in $1 billion takeover talks
21 1 15 22 French supermarkets face food crisis following cyberattack on refrigerators
20 0 24 32 Meta Platforms launches Muse Wonder AI model
18 1 15 18 Microsoft launches third-generation Maia AI chip
16 1 5 16 ICICI Prudential AMC declares record dividend and reports Q2 profit rise
15 2 6 20 HCL Tech Q2 profit rises 21% year over year
15 0 38 97 Saudi Arabia's Riyadh airport under Houthi attacks
Velocity alone is not enough: one site publishing a long article produces many statements, so an entry can come from a single outlet. Filter on breadth:
import requests
API = "https://finance-pulse.p.rapidapi.com"
HEADERS = {"X-RapidAPI-Key": "YOUR_RAPIDAPI_KEY", "X-RapidAPI-Host": "finance-pulse.p.rapidapi.com"}
topics = requests.get(f"{API}/v2/topics", headers=HEADERS,
params={"sort": "velocity", "min_statements": 5, "limit": 20}).json()["data"]
developing = [t for t in topics if t["sources"] >= 5 and t["statements_4h"] >= 3]
for t in developing:
print(f"{t['sources']:>3} outlets {t['statements_4h']:>2} in 4h {t['name']}")
16 outlets 4 in 4h LVMH's luxury-goods sales slump in China
11 outlets 10 in 4h Visa approved to establish bank card clearing institution in China
6 outlets 6 in 4h Reliance Jio Platforms IPO
7 outlets 4 in 4h Broadcom CEO likens AI boom to casino
What the market is picking up right now
/v2/trending is the API's own selection, ranked by distinct outlets: a live list of what is most reported
now, and on every topic the labels of the earlier time slots in which it trended. kind=live returns just the live
list:
#1 3 outlets on the live list (11 in total) Visa approved to establish bank card clearing institution in China
#2 3 outlets on the live list (4 in total) IMF warns of global economic risks
#3 3 outlets on the live list (3 in total) B&Q owner names Sebastian Gundel as incoming CEO
#4 3 outlets on the live list (6 in total) Blur NFT marketplace shuts down website
#5 3 outlets on the live list (6 in total) Poundland management in talks to lead bid to save 11,000 jobs
#6 2 outlets on the live list (6 in total) ASML Holding N.V. revenue growth
#7 2 outlets on the live list (2 in total) Emissary Partners hires senior advisors
#8 2 outlets on the live list (4 in total) US stock index futures drop
#9 2 outlets on the live list (8 in total) German retail chain Galeria faces fourth insolvency in six years
#10 2 outlets on the live list (3 in total) BMW second-quarter revenue and earnings drop
Each item also carries its newest statement, with its image when the article had one, so one request is enough to render a front page. fintopic.news is built this way.
Follow one story
Take the top developing story and look inside it:
story = requests.get(f"{API}/v2/topics/{developing[0]['id']}", headers=HEADERS).json()["data"]
print(story["name"], "|", story["statements"], "statements,", story["sources"], "outlets")
print("sentiment", story["sentiment"], "| source mix", story["source_mix"])
Robert Hall and Michael Woodford win Nobel Prize in Economics | 58 statements, 42 outlets
sentiment {'bullish': 0, 'bearish': 0, 'neutral': 58} | source mix {'news': 58}
Then its life, hour by hour, from /v2/series?topic_id=…&interval=hour:
Data table
| hour (UTC) | statements | sources |
|---|---|---|
| 2026-10-12T05:00 | 1 | 1 |
| 2026-10-12T09:00 | 8 | 7 |
| 2026-10-12T10:00 | 26 | 21 |
| 2026-10-12T11:00 | 5 | 5 |
| 2026-10-12T12:00 | 3 | 3 |
| 2026-10-12T13:00 | 2 | 1 |
| 2026-10-12T14:00 | 4 | 4 |
| 2026-10-12T17:00 | 2 | 2 |
| 2026-10-12T19:00 | 1 | 1 |
| 2026-10-12T20:00 | 2 | 2 |
| 2026-10-12T21:00 | 2 | 1 |
| 2026-10-12T23:00 | 1 | 1 |
| 2026-10-13T01:00 | 1 | 1 |
The shape is typical: a first report, a quiet stretch, then a burst as more outlets pick it up. The burst, in
outlets rather than statements, is what to alert on: wire /v2/feed?topic_id=… (see
the feed guide) to the topics that pass your breadth filter.
Storylines: one company's stories in order
A topic is one development. A company's story is usually several: the announcement, the reaction, the analyst notes,
the follow-up a week later. A storyline groups the topics about one symbol that continue each other.
/v2/storylines?symbol= lists them; min_topics=2 leaves out the storylines that hold a single topic:
curl -G https://finance-pulse.p.rapidapi.com/v2/storylines \
--data-urlencode "symbol=MSFT" --data-urlencode "min_topics=2" --data-urlencode "sort=mentions" \
-H "X-RapidAPI-Key: YOUR_RAPIDAPI_KEY" -H "X-RapidAPI-Host: finance-pulse.p.rapidapi.com"
topics mentions outlets first seen .. last seen storyline
9 48 41 10-08 .. 10-10 US government suspends Microsoft and Adobe from green card program
8 31 26 10-10 .. 10-13 Microsoft warns AI faces less regulation than milk
3 25 18 08-18 .. 10-13 Microsoft launches third-generation Maia AI chip
2 22 24 09-19 .. 10-12 US suspends Microsoft from green card program
3 21 5 08-13 .. 10-12 Microsoft Azure revenue crosses $100 billion annual pace
4 17 12 08-13 .. 10-12 Microsoft unifying Copilot AI tools into a super app
3 15 11 08-13 .. 10-12 Microsoft posts trillion-dollar quarter
4 11 10 10-07 .. 10-08 Microsoft and Nvidia team up on AI laptops
Each row carries up to five of its topics (top_members); /v2/storylines/{id} lists all of them, oldest first, with
the newest statements:
line = requests.get(f"{API}/v2/storylines", headers=HEADERS,
params={"symbol": "MSFT", "min_topics": 2, "sort": "mentions"}).json()["data"][0]
detail = requests.get(f"{API}/v2/storylines/{line['id']}", headers=HEADERS).json()["data"]
for t in detail["members"]:
print(t["created_at"][:10], f"{t['mentions']:>3} of {t['statements']:>3} statements name MSFT ", t["name"])
for other in detail["related"]:
print("also under", other["symbol"], "-", other["name"], f"({other['shared_topics']} shared topics)")
2026-10-08 24 of 25 statements name MSFT US government suspends Microsoft and Adobe from green card program
2026-10-08 5 of 5 statements name MSFT Microsoft accused of fraud by JD Vance
2026-10-08 1 of 2 statements name MSFT U.S. Labor Secretary suspends Microsoft, Adobe, Cognizant, Infosys, Tata Consultancy Services, Wipro, HCL Technologies, and Capgemini from PERM
2026-10-08 7 of 7 statements name MSFT US bans Microsoft from sponsoring foreign workers
2026-10-08 3 of 3 statements name MSFT Microsoft accused of firing 6,000 Americans for foreign indentured servants
2026-10-08 3 of 4 statements name MSFT US immigration crackdown targets Microsoft and tech firms
2026-10-08 2 of 3 statements name MSFT US suspends Microsoft and Indian tech firms from visa program
2026-10-09 1 of 2 statements name MSFT US tech firms blocked from green cards
2026-10-09 2 of 2 statements name MSFT Microsoft faces US visa ban
also under ADBE - US government suspends Microsoft and Adobe from green card program (2 shared topics)
also under WIPRO.NS - U.S. Labor Secretary suspends Microsoft, Adobe, Cognizant, Infosys, Tata Consultancy Services, Wipro, HCL Technologies, and Capgemini from PERM (1 shared topics)
also under INFY - US suspends PERM for Indian IT majors (1 shared topics)
also under TCS.NS - U.S. Labor Secretary suspends Microsoft, Adobe, Cognizant, Infosys, Tata Consultancy Services, Wipro, HCL Technologies, and Capgemini from PERM (1 shared topics)
How to read the numbers:
statementscounts every statement of the member topics;mentionsonly those that name the symbol, the same meaning as on/v2/symbols. Rank bymentions: a member topic need not be mostly about the company, sostatementscan overstate the company's part. The two bases do not reconcile:statementsandsourcescover everything/v2/statements?storyline_id=…returns,mentionsand the three sentiment counts only what remains when you add&symbol=MSFT.bullish/bearish/neutralare the labels of the mentions, as counts. There is no score for a storyline: it holds an event and the reactions to it, so its tone is a path, not a level. For organisations, rates and currencies the labels have no single referent.first_seen/last_seenare the oldest and newest statement of it within your plan's history, so never beforesnapshot.window_start. They are not the story's beginning and end.relatedlists storylines of other symbols that share topics with this one (the five that share the most): the same news seen from the other company. It is an overlap, not a claim that the two are one story./v2/metacarriesstorylinesandstorylines_multi_topic(how many the API holds, and how many of them group more than one topic: bookkeeping, not a count of stories). Whenstoryline_methodthere changes, storyline ids and memberships you stored may no longer be valid: fetch them again. A stored storyline id can also stop working when its company's code changes (a renamed ticker, for example) or when the storyline is merged into another: fetch/v2/storylines?symbol=again. Memberships from beforestorylines_live_sincehave noplaced_at.- A topic says where it sits:
"storylines": [{"id": …, "symbol": "MSFT", "topics": 7, "placed_at": …}].[]means it is in no storyline, even ifsymbolslists some.nullmeans it has not been placed yet."topics": 1only says no other member is within your plan's history, not that the story is new. A topic'srevisiondoes not change when it is placed: readstorylinesitself.storyline_id=filters/v2/topicsand/v2/statements.
Run this yourself: get an API key on RapidAPI ↗.