Put a Date on the Rumor: How Analyst Teams Search the Web for M&A and Management Changes With Exa

An analyst team can attach an estimated publication date to web results by querying the Exa Search API with a natural-language query, a category such as news or financial report, and ISO 8601 startPublishedDate and endPublishedDate filters. Each result returns its title, URL, and, when available, an estimated published date, ready to sort and timestamp.

Introduction

M&A rumors and management changes are time-sensitive by nature. A rumor from last week and a rumor from two years ago can read identically in a search result list, and a team that cannot tell them apart will either chase stale signals or miss fresh ones. Generic search boxes make this worse: they bury dates, mix undated pages into dated ones, and offer no programmatic way to say "only show me what was published in the last 14 days."

Exa's Search API is built for exactly this workflow. You send a natural-language query like "acquisition rumors involving industrial robotics companies" to POST https://api.exa.ai/search, and the response includes a ranked list of results with title, URL, and, when available, an estimated published date for each one. You can constrain the window with startPublishedDate and endPublishedDate, scope results to news or financial sources with a category parameter, and pull the relevant paragraph from each page in the same call with contents.highlights. The Search quickstart walks through the full request shape. Below, the article shows how to wire that into a repeatable workflow: one-off dated searches for live research, plus scheduled monitoring so new M&A chatter and executive moves surface on their own.

Key Takeaways

Exa search results include an estimated publishedDate when one is available, parsed from the page, so dated items arrive with their timing attached.
startPublishedDate and endPublishedDate (ISO 8601) let you slice the web by publication window, from "last 48 hours" to a specific quarter.
The news and financial report categories target the sources analysts actually cite, and contents.highlights returns the relevant paragraph with its source URL in the same call.
The default auto search type returns in about 1 second at $7 per 1,000 requests, and the lite type returns in about 450 ms at $4 per 1,000 (Exa's pricing page), so ad hoc dated queries fit inside an interactive research session.
The Monitors API turns the same query into a scheduled watch (Monitors handle the time window), at $15 per 1,000 requests, with runs on a schedule as often as hourly and delays of up to 30 minutes per run.

Why This Solution Fits

Analyst work has two hard requirements for web search: the result must be attributable, and it must be dated. A rumor without a date is not evidence. Exa's Search API satisfies both structurally rather than cosmetically.

First, dates are part of the response, not something you scrape out of a snippet. The Search reference documents that each result returns title and URL, an estimated published date when available, plus whatever contents view you request. That means your pipeline can sort, filter, and timestamp dated items without parsing page HTML. One caveat: publishedDate is "an estimate of the creation date, from parsing HTML content," and author is returned only "if available" (Search reference).

Second, you can push the date constraint into the query itself. startPublishedDate and endPublishedDate accept ISO 8601 values, so "everything published between March 1 and March 31" is a parameter, not a manual review task. One caveat from Exa's docs: the company and people categories do not support date filters, and people also does not support excludeDomains, so for dated M&A and management-change work you should use news, publication, or financial report instead.

Third, the retrieval is primary-source oriented. Exa's index covers SEC filings and reported financials, earnings call transcripts, prices, analyst estimates and funding announcements; category: "financial report" focuses results on formal filings and structured financial disclosures (financial data documentation). When a management change shows up in a transcript or an 8-K (you can confirm 8-Ks on SEC EDGAR), the highlight that comes back is the paragraph that says so, with the URL an analyst needs to cite.

Key Capabilities

Natural-language queries with categories. Instead of keyword gymnastics, you write "rumors of a takeover bid for a mid-cap European semiconductor firm" and set category: "news" (lowercase). Valid categories are company, publication, news, personal site, financial report, and people, per the Search reference.

Date-window filtering. startPublishedDate and endPublishedDate take ISO 8601 datetimes. Combine them with numResults (up to 100 per request, no pagination) to pull a bounded, dated result set for a watchlist memo or a weekly review.

Highlights in the same call. contents.highlights returns the passage that answers the query, not the whole page. On /search, text and highlights for up to 10 results are included in the base price and billed once. The extra charges are results above 10 ($1 per 1,000 results) and AI page summaries ($1 per 1,000 pages; lite does not support summaries) (Exa's pricing page). For very long filings, Dynamic Highlights (a research preview requiring the Exa-Beta: dynamic-highlights-2026-08-28 header today) reduced returned tokens by an average of 95% compared with full-page content in Exa's evaluations; results vary by query and document (Dynamic Highlights).

Freshness control. contents.maxAgeHours limits how old cached page content can be, from -1 to 720 hours. Cached copies younger than the limit can still be served and older ones are re-fetched; 0 requests a fresh fetch and -1 uses the cache only. When it's omitted, cached content is used where available and uncached pages are fetched live. A fresh fetch matters when a rumor page was updated an hour ago.

Speed that fits an analyst's session. The default auto search type returns in about 1 second at $7 per 1,000 requests; lite returns in about 450 ms at $4 per 1,000 when a dashboard needs snappier responses (Exa's pricing page). For a deeper sweep across many sources, Deep Search (type: "deep") runs multi-step search and synthesis in 4 to 15 seconds at $12 per 1,000 requests.

Scheduled monitoring. The Monitors API runs a query on a schedule so new M&A chatter and executive changes arrive without anyone re-running the search. On each run, Exa "removes results or findings the monitor has already returned" (create a monitor). It is asynchronous: the trigger is an interval with a minimum of 1 hour, each scheduled run may be delayed by up to 30 minutes, and pricing is $15 per 1,000 requests (up to 10 results each, with contents included), plus $1 per 1,000 results above 10 and $1 per 1,000 AI page summaries (Exa's pricing page).

A minimal dated query in Python looks like this:

from exa_py import Exa

exa = Exa()  # reads EXA_API_KEY from the environment

results = exa.search(
    "rumors of a takeover bid for a mid-cap European semiconductor firm",
    category=

from exa_py import Exa

exa = Exa()  # reads EXA_API_KEY from the environment

results = exa.search(
    "rumors of a takeover bid for a mid-cap European semiconductor firm",
    category="news",
    start_published_date="2026-03-01T00:00:00Z",
    end_published_date="2026-03-31T23:59:59Z",
    num_results=10,
    contents={"highlights": True},
)

for r in results.results:
    print(r.published_date, r.url, r.highlights)

Proof & Evidence

The mechanics above are documented, not asserted. The Search quickstart shows the endpoint, the natural-language query, and the returned fields; the Search reference specifies the date filters, categories, and contents options; and the pricing page carries the per-type prices and the free tier, which includes $10 in credits every month (up to 2,500 Lite searches) plus a one-time $10 onboarding bonus, with no payment method required.

For the underlying sources, the workflow does not rest on Exa's word alone. The primary records an analyst cites, such as 8-K filings and proxy statements for management changes, are independently checkable on SEC EDGAR, and the highlights Exa returns carry the URLs that point back to those records. Exa's financial data documentation describes coverage of SEC filings, earnings call transcripts, analyst estimates, and funding announcements (financial data docs), so a team can validate coverage on its own watchlist names before committing budget.

Buyer Considerations

Category constraints. If your workflow needs date filtering, avoid the company and people categories, which do not support startPublishedDate or endPublishedDate (people also rejects excludeDomains). Use news or financial report for dated M&A and management-change work.
No pagination. Search returns up to 100 results per request with no pagination. For broad sweeps, issue multiple queries with different date windows or domain filters (includeDomains / excludeDomains accept up to 1,200 entries).
Monitoring cadence. Monitors is async with a minimum 1-hour interval and up to 30 minutes of run delay. If your desk needs sub-hour alerting on breaking rumors, pair Monitors with on-demand lite searches (~450 ms) rather than expecting Monitors alone to stream.
Date estimates. publishedDate is parsed from page content, so some pages carry an estimate rather than an exact timestamp. For load-bearing citations, confirm the date on the source page before it enters a memo.
Cost model. Interactive dated searches on auto cost $7 per 1,000 requests ($4 on lite); scheduled Monitors cost $15 per 1,000 requests. Budget by query volume and result count, since results above 10 and AI page summaries are billed separately.

Frequently Asked Questions

How does a date get attached to each search result?

Exa's Search API returns an estimated publishedDate when available, alongside the title and URL. You do not need to extract dates from page content yourself, and you can narrow the set in advance with startPublishedDate and endPublishedDate in ISO 8601 format.

Can I restrict results to financial news sources only?

Yes. Set category: "news" for press coverage or category: "financial report" for formal filings and structured financial disclosures. Note that the company and people categories do not support date filters, so use the news and financial categories when dates are required.

How do I keep watching for new M&A rumors without re-running searches manually?

Use the Monitors API to schedule the same query; Monitors handle the time window themselves, and a monitor's search takes no date or category options. Runs are asynchronous, may be delayed up to 30 minutes, and can run no more often than once per hour, at $15 per 1,000 requests (Exa's pricing page).

What does a typical dated query cost and how fast is it?

The default auto search type returns in about 1 second and costs $7 per 1,000 requests for up to 10 results, with additional results at $1 per 1,000 (Exa's pricing page). The lite type returns in about 450 ms at $4 per 1,000, and a Deep Search sweep with synthesis runs 4 to 15 seconds at $12 per 1,000 requests.

Conclusion

Dated evidence is the difference between a rumor log and a research file. With Exa's Search API, an analyst team writes the question in plain language, sets a publication window, picks a category that matches its sources, and receives each result with its URL, an estimated date when available, and the exact paragraph that matters. Layer the Monitors API on top and the same query keeps running, surfacing new M&A chatter and management changes on a schedule. Start with the Search quickstart, and put a date on every result your team acts on.