Last updated: 24 August 2026
marlow is a browser extension that checks the claims and citations in web articles. When you visit a web page, the extension may send the page text to our servers for analysis. The analysis checks whether quoted experts exist, whether cited studies are real, and whether claims match their supposed sources.
Page content: When marlow scans a page, the text content of that page is sent to our analysis servers. This is necessary to perform the trust analysis. We do not store the full text of pages you visit — after analysis the full text is discarded. We do retain a short excerpt (up to 500 characters) alongside the scan result, used to check analysis quality and to review disputed findings. Scan records are not linked to your name, account, or any device identifier (see Data retention).
Documents you choose to scan: marlow can also analyse documents you open or drag into the extension (for example PDF or Word files). When you scan a document, its text is sent to our analysis servers in the same way as web-page text — processed to perform the trust analysis and then discarded, with the same short-excerpt retention as page scans. Documents are read on your device and only their text is sent. We never access files you have not explicitly chosen to scan.
Page URL and domain: The URL and domain of scanned pages are recorded to build domain-level trust profiles. This helps marlow provide faster, more accurate results over time.
Scan results: The analysis results (the checks that were run, the evidence found, and the state each claim landed in), the page URL, and a short excerpt (up to 500 characters) of the page are stored to operate and improve the checking feature, including the domain-level results the extension shows you, and to review disputed findings. These records carry no account, contact, or device identifier — they are not linked to you and do not contain personally identifiable information.
Usage data: We track aggregate usage metrics such as the number of scans performed. We do not track which specific pages you visit beyond the scanning functionality.
Scan rate limiting: marlow limits how many scans a free user can run each day. To enforce this limit fairly per installation rather than per device or per IP, each scan request carries a pseudonymous installation identifier — a random string generated on your device when the extension is installed. The identifier is not linked to your name, email, account, or any personally identifiable information. The identifier is stored locally in your browser and on our infrastructure for rate-counting purposes. No account creation is required.
Page text is sent to Anthropic for analysis. Anthropic's API processes the text and returns the results of those checks. Anthropic's API usage is subject to Anthropic's privacy policy.
Analysis results are stored on Cloudflare infrastructure. Cloudflare's data processing is subject to Cloudflare's privacy policy.
To check claims in the articles you scan, marlow also cross-references short fragments of the article text against a small set of public third-party sources. This is how we verify whether quoted experts exist, cited studies are real, and named events appear in independent coverage.
What is sent: Entity names (people, organisations, places, works), short quote fragments, and citation identifiers such as DOIs — all extracted from the article text you are already scanning. We do not send the full article text, the URL of the page you are viewing, or any identifier associated with you or your device.
Categories of third parties we query, and why:
The specific third-party services used within each category are maintained internally and provided on request to data-subject inquiries (privacy@marlowverify.com).
Caching: Lookup results are cached on our infrastructure. Cache entries are not associated with any user or device.
Each third party receives anonymous HTTP requests and handles them under its own privacy policy.
marlow includes an optional "Was this right?" link in the scan panel. If you choose to use it, your response is sent to our servers so we can improve the checks over time. Nothing is sent unless you actively submit feedback.
What is sent when you submit feedback:
We do not send the article URL, the article content, or any identifier that could be used to contact you. Feedback responses are stored on Cloudflare infrastructure and retained and used to improve checking accuracy. You can stop all feedback collection at any time by uninstalling the extension.
You can uninstall the marlow extension at any time, which immediately stops all data collection. Because scan records carry no identifier linking them to you, we cannot locate them by your identity; if you want the records for a specific page URL or domain removed, send that URL or domain to privacy@marlowverify.com and we will delete the matching records. You can also ask which scan records exist for a given URL or domain, and ask us to correct a record you believe is wrong, using the same address.
Data leaves our infrastructure to the following named processors, each governed by its own privacy policy:
In addition, short article-extracted fragments (entity names, quote fragments, citation identifiers — never the full article text or any device identifier) are sent to public knowledge databases, scholarly and statistical data sources, retraction registries, fact-check aggregators, and news archives for verification purposes, as described in the Knowledge-layer verification section above. The current roster of specific services within those categories is available on request.
Our use of information received through the marlow extension, including any information received from Google APIs, adheres to the Chrome Web Store User Data Policy, including the Limited Use requirements. We collect only the data needed for the single purpose of the extension — checking the claims and citations in the pages and documents you choose to scan — and for related operational purposes such as maintaining, securing and improving that feature. We do not sell user data, and we do not use it for advertising or for any purpose unrelated to that single purpose.
marlow is operated from Melbourne, Australia. The processors named above operate infrastructure outside Australia, including in the United States, so page text sent for analysis and the records we store may be processed and held overseas.
Traffic between the extension, our servers and our processors is encrypted in transit using HTTPS. Access to stored records is restricted to the operator of the service. No system is perfectly secure. If a breach occurs that is likely to result in serious harm, we will notify affected users and the Office of the Australian Information Commissioner in line with the Notifiable Data Breaches scheme.
If you have a privacy complaint, email privacy@marlowverify.com and we will respond. If you are not satisfied with our response, you can refer the matter to the Office of the Australian Information Commissioner at oaic.gov.au.
We may update this privacy policy from time to time. Changes will be posted on this page with an updated date. Continued use of the extension after changes constitutes acceptance of the updated policy.
For privacy-related questions, contact: privacy@marlowverify.com