Key Takeaways
- Filing language is carried forward, so the changed sentences are a short, deliberate list rather than noise.
- Compare one section at a time. A whole-filing comparison returns thousands of differences and hides the useful ones.
- Added risk factors and removed risk factors are the two lists worth reading first.
- The ignore-numbers option separates a reworded disclosure from one that only restated its figures.
- Similarity is a text measurement, not a verdict. One added sentence can matter more than a full rewrite.
- Both versions stay in your browser. The comparison makes no network request.
What Is a Filing Difference Viewer?
A filing difference viewer aligns two versions of the same text and reports what moved between them. It is the same idea as a redline in a contract negotiation, applied to disclosure documents that are revised annually and rarely rewritten.
The reason it works on SEC filings specifically is that filers reuse language. A risk factor drafted in 2019 is still in the 2026 filing, word for word, unless someone decided otherwise. That stability turns every change into a signal about intent: legal and management reviewed the section, and this is what they chose to alter.
Reading two filings side by side does not produce the same result. The eye skips over unchanged paragraphs and stops on ones it half-recognizes, which is the opposite of what is needed. Alignment is mechanical work, and mechanical work is what a tool is for.
SEC Filing Difference Viewer
Paste the older version on the left and the newer version on the right. One section at a time gives the most readable result.
All processing runs in your browser. Neither version is sent to Swoopr Investment's servers, stored, or logged.
How the Comparison Works
- Normalize. Non-breaking spaces, soft hyphens, zero-width characters, typographic quotes and dashes, and lines holding only a page number are all folded away. These are copy-paste artifacts, and without this step they would each register as a change.
- Split into units. Sentences by default, paragraphs on request. Sentence splitting keeps abbreviations intact, so "U.S." and "Nasdaq, Inc." do not fragment into false differences.
- Align. A longest-common-subsequence alignment finds the largest set of units present in both versions, in order. Everything outside that set is an addition or a removal.
- Measure. Similarity is twice the matched unit count divided by the total unit count across both versions. A value of 1 means the two versions are identical after normalization.
similarity = (2 x matched units) / (previous units + current units)
Each version is capped at 1,200 units. The alignment is quadratic in the unit count, and that cap is what keeps a pasted full annual report from locking the browser tab. When the cap is reached the result says so rather than quietly comparing part of the text.
Worked Example: A Risk Factor Added
Hypothetical, and small enough to check by hand. Last year's section held three sentences. This year's holds the same three plus one new one.
| Measure | Value | Where it comes from |
|---|---|---|
| Earlier version units | 3 | Three sentences. |
| Later version units | 4 | Three carried forward, one added. |
| Matched | 3 | Every earlier sentence survives unchanged. |
| Added | 1 | The new sentence. |
| Removed | 0 | Nothing dropped. |
| Similarity | 85.7% | (2 x 3) / (3 + 4) = 6/7. |
| Change rate | 14.3% | (1 + 0) / (3 + 4) = 1/7. |
The percentages are almost beside the point here. The output that matters is one line long: a cybersecurity risk factor was added. That is the fact worth carrying into a research note, with the filing link attached.
A Year-Over-Year Reading Workflow
- Use the SEC Filing Section Navigator on both filings to see which sections changed size. A section that grew or shrank materially is where to look.
- Copy that Item from each filing and compare it here, one section at a time.
- Read the added list first, then the removed list. Both are short and both are deliberate.
- Turn on ignore-numbers and compare again. Sentences that were unchanged the first time and are now identical were pure figure updates; the remaining differences are language.
- Record what you found in the Investment Research Notebook with the filing URL, so the finding keeps its source.
- If the change concerns reported figures, check the tagged data in the XBRL Financial Statement Explorer rather than trusting a number read out of prose.
What This Tool Does Not Do
- It does not rank changes by importance. Every added sentence is listed equally. Deciding which one matters is the reader's job and cannot be automated honestly.
- It compares text, not meaning. A paragraph rewritten to say the same thing registers as a removal plus an addition.
- It does not detect moved text. A section reordered without editing shows as changes, because position is part of the alignment.
- It does not fetch filings. Both versions are pasted by you.
- It is not a legal redline. There is no character-level markup and no attempt to reproduce a law firm's comparison conventions.
- Nothing here is investment advice. A changed risk factor is information, not a recommendation.
Frequently Asked Questions
How do I compare two SEC filings?
Compare one section at a time rather than two whole documents. Open both filings, copy the same Item from each (Item 1A Risk Factors from last year and from this year, for example), paste the older version on the left and the newer on the right, and run the comparison. Section-level comparison is what produces a readable result: a whole-filing diff reports thousands of changes, most of them dates and figures, and buries the handful that matter.
What is a risk factor diff and why does it matter?
A risk factor diff is a sentence-level comparison of Item 1A across two filings. It matters because risk factors are drafted by counsel, carried forward largely unchanged year to year, and edited deliberately. A newly added risk factor was added on purpose. A removed one was removed on purpose. Those two lists are usually short and are the most direct statement a company makes about what changed in its own view of its business.
Does a large percentage change mean something is wrong?
No. The similarity figure measures text, and text moves for reasons that have nothing to do with the business: a new drafting team, a reformatted document, a reorganized section order, or a paste that picked up page headers. A high change rate means read the changes, not that something is wrong. Equally, a very low change rate can hide a single added sentence that matters more than everything else on the page.
What does the ignore-numbers option do?
It replaces every number with a placeholder before comparing, so two sentences that differ only in their figures register as unchanged. That separates two different kinds of edit: a disclosure that was reworded, and a disclosure that kept its wording and restated its figures. Turn it on to find genuine language changes; leave it off when the figures themselves are what you are checking.
Is my filing text sent anywhere?
No. The comparison runs entirely in your browser. Neither version is uploaded, stored, or logged, and no analytics event carries any part of the text. The tool makes no network request at all while comparing.
References
Disclosure-requirement claims are sourced to SEC form documents and rulemaking releases, verified in August 2026. The worked example is hypothetical and arithmetic only; no statistic on this page comes from an external dataset.
- SEC: Form 10-K: the annual report form, and the source for the Item structure that makes section-by-section comparison possible.
- SEC: Modernization of Regulation S-K Items 101, 103, and 105: the amendments to the risk factor requirement, including the summary requirement for lengthy risk factor sections.
- SEC: Cybersecurity Risk Management, Strategy, Governance, and Incident Disclosure: an example of a rulemaking that produced new disclosure language across many filers at once, which a year-over-year comparison surfaces immediately.
- SEC: Accessing EDGAR Data: how to retrieve the filings whose sections you compare here, and the fair-access expectations that apply.
Jurisdiction: United States. Last reviewed by the Swoopr Editorial Team in August 2026. This page is educational and is not personalized investment advice.