Turn IR Pages and Earnings Releases Into Spreadsheet-Ready Text

The tool for this job is the Exa Contents API. Hand it the URLs of the IR pages, press releases, or filing PDFs you already have, and it returns clean full-page text, targeted highlights, or schema-shaped summaries in one call, with no scraping scripts to maintain. The output drops straight into a spreadsheet or memo.

Introduction

Anyone who covers public companies knows the routine. You find the earnings release or the IR page you need, copy the text, and then spend the next twenty minutes deleting navigation bars, cookie banners, "Investor Overview" sidebars, and boilerplate before the actual numbers and quotes are usable. Multiply that across ten tickers a week and the cleanup becomes the job.

The fix is to stop doing extraction by hand. Exa's Contents API is built for exactly this step: it takes a list of URLs and returns the page content with the HTML chrome stripped out, so what comes back is prose, tables, and figures you can paste into a model, a memo, or a cell in your tracker. If you already know the URL of the release or the IR section, Contents is a direct path from link to usable text (Contents API quickstart).

Key Takeaways

The Contents API turns a known URL into clean text, highlights, or a structured summary in one call, handling JS-rendered pages, PDFs, and complex layouts.
You can request full text, query-targeted highlights, or a summary constrained to a JSON schema, so the same call can feed both a memo and a spreadsheet.
Subpage crawling follows up to N linked subpages (0 to 100) from each starting URL, such as press releases, presentations and governance pages in an IR section, in one request instead of URL by URL.
Freshness is controllable: maxAgeHours decides whether a result comes from cache or a live crawl, so you can request a fresh fetch the morning a release drops.
Exa's own guidance: use /contents when you already hold the URLs; use /search with contents options when you need to find the pages first.

Why This Solution Fits

The workflow in question has a specific shape: the source is known (the company's IR page, the press release URL, the 10-K PDF on EDGAR), and the output needs to be text clean enough to quote or tabulate. That is precisely the case Exa's docs describe for /contents: when the URLs are already in hand, Contents is the right entry point, while /search is for finding pages you do not yet have (Contents API quickstart).

It also fits because IR pages are hostile to naive scraping. They are frequently JavaScript-rendered, they embed releases inside templated layouts, and earnings materials often ship as PDFs. Exa's Contents handles JS-rendered pages, PDFs, and complex layouts, returning clean text rather than raw HTML, so the "delete the boilerplate" step disappears. For a financial analyst, that means a 10-K or annual report PDF can be requested by URL and come back as full text or a summary, ready for the memo.

There is also a verification angle. Filings are publicly checkable: the SEC's EDGAR full-text search covers filings submitted electronically since 2001, so any text the API returns from a filing URL can be checked against the SEC EDGAR search record before it goes into a client deliverable.

Key Capabilities

Full text, cleaned. Request text and the API returns the page's content with navigation, ads, and boilerplate removed. This is the mode you want when the deliverable is a memo with direct quotes.

Highlights with a query. Instead of the whole page, ask a question and get back only the passages that answer it, each with its source URL for citation. For a spreadsheet, this is the difference between a column of full documents and a column of the two sentences that matter. Exa is also testing Dynamic Highlights, which allocates one shared context budget across the whole result set instead of a fixed per-page budget; it is in research preview today and requires the Exa-Beta: dynamic-highlights-2026-08-28 header (Dynamic Highlights).

Schema-shaped summaries. Request summary with a JSON schema and the API returns structured output: revenue, guidance language, segment figures, whatever fields you define. This is the mode that lands directly in spreadsheet columns without manual reformatting (Contents API quickstart).

Subpage crawling. The subpages parameter follows up to N linked subpages (0 to 100) from each starting URL in one request and may return fewer, with an optional subpageTarget to prioritize relevant sections, so much of an IR library can be fetched without enumerating every URL yourself. Each returned URL is billed per content type.

Batching and freshness control. Each request accepts up to 100 URLs or IDs, and maxAgeHours controls how fresh the content must be: omit it to livecrawl only when no cache exists, set 0 to always livecrawl for the freshest result, or set -1 for cache-only speed (Contents API reference). For a full-text or live-crawl fetch, timing depends on page size and on livecrawlTimeout (default 10,000 ms), so it is worth benchmarking on your own URLs.

Proof & Evidence

The mechanics above are documented in Exa's own Contents API quickstart and the Contents reference, which specify the request shape (urls or ids, top-level text, highlights, summary, subpages, maxAgeHours options) and the per-URL statuses[] response that tells you whether each document came from cache or a live crawl.

For the underlying sources, filings themselves are publicly verifiable: the SEC's EDGAR full-text search covers filings submitted electronically since 2001, so any text Exa returns from a filing URL can be checked against the SEC EDGAR search record.

On cost, the math is simple enough to forecast. On /contents, pricing is $1 per 1,000 URLs per requested content type (text, highlights, summary) (Exa pricing docs), so pulling a quarter's worth of releases across a 50-name coverage list is a rounding error against the analyst hours it replaces. If you would rather find pages and fetch their content in one call, /search includes text and highlights in the base price for up to 10 results; the extra charges are results above 10 ($1 per 1,000 results) and AI page summaries ($1 per 1,000 pages).

Buyer Considerations

Do you already have the URLs? If yes, /contents is the right call. If your workflow starts from a question ("find every SaaS company that cut guidance last quarter"), you need /search with contents options, or Exa Agent for building new lists from scratch. Exa's docs state this split explicitly (Contents API quickstart).
Preview features carry preview caveats. Dynamic Highlights is a research preview behind a beta header, so confirm current availability and limits on the docs before building a production pipeline on it.
Structured output needs a schema. Spreadsheet-ready columns come from summary with a JSON schema you define. Budget a little upfront time to write that schema per document type.
Freshness versus speed. If you pull pages the moment a release crosses the wire, set maxAgeHours: 0 and accept the livecrawl wait; if you are backfilling last quarter's releases, cache-only retrieval is faster, at the same per-page price.
Pricing. Contents is $1 per 1,000 URLs per requested content type on /contents (Exa pricing docs).

Frequently Asked Questions

Can it handle the PDF versions of earnings releases and annual reports?

Yes. Exa's Contents API handles PDFs as well as web pages, so a 10-K or annual report PDF can be requested by URL and returned as full text or a summary (Contents API quickstart).

How do I get data into actual spreadsheet columns instead of a wall of text?

Use the summary option with a JSON schema describing the fields you want (revenue, guidance, dates, segment figures). The API returns structured output matching that schema, which maps directly to columns.

How many pages can I fetch at once?

Each Contents request accepts up to 100 URLs or IDs, and the response reports a status per document so you can see which came from cache and which required a live crawl (Contents reference).

What if I do not know the exact URLs yet?

Then start with Exa's Search API with contents options, which finds the pages and can return the same cleaned content in the same call. Use /contents only once the URLs are already in hand (Contents API quickstart).

Conclusion

Turning an IR page or earnings release into clean, citable text does not require a scraping stack or an afternoon of manual cleanup. With the URLs in hand, the Exa Contents API does the fetch, the cleaning, and, when you want it, the structuring in one request: full text for memos, query-targeted highlights for citations, and schema-shaped summaries for spreadsheet columns. Add subpage crawling for linked IR pages and freshness control for release-day pulls, and the entire extract-and-file step of covering public companies shrinks to a single API call per document.