Checklist

What tattle.api.osano.com sent after Reject on four Osano stores

On three Osano stores, Chrome sent a tattle.api.osano.com report about 30 seconds after load, after Reject too. Reject went to consent.api.osano.com.

In brief

tattle.api.osano.com is the Osano host that receives reports about scripts, cookies, and storage keys a site's Osano setup has not classified yet. We tested four Shopify stores that use Osano, Stores K, R, AP, and AS. On three of them, the browser sent a tattle report about 30 seconds after a page load, before and after Reject. None of the report entries we could read held a consent choice. The Reject choice went separately to consent.api.osano.com/record. In our HTTP/1.1 rerun, 72 of 75 tattle attempts got a 204 on the preflight and a 204 on the POST, including 13 of the 15 sent after Reject. None got a 403. Technical observation, not legal advice.

What Osano documents about tattle

Osano's December 2019 release notes describe "tattles" from its Consent Manager. The script groups them, sends one request on a schedule, stores them in the browser so the same visitor does not report the same item twice, and throttles them until the site owner clears the queue of uncategorized scripts.

Osano's docs tie these reports to Discoveries. Osano reports the full URL of scripts that try to load and the cookies it finds, and it can also report Local Storage items. In Listener mode it blocks nothing. In Permissive mode it reports uncategorized scripts and lets them run. In Strict mode it blocks them until someone classifies them. We did not check which mode any of the four stores uses.

Osano's CSP guide lists tattle.api.osano.com next to consent.api.osano.com, cmp.osano.com, and disclosure.api.osano.com as subdomains to allow. Osano's cookie page says osano_consentmanager holds the consent categories and osano_consentmanager_uuid is the visitor's ID for consent lookups.

How we tested

Stores K and R are the two Osano stores from our Postscript after Reject test, and Stores AP and AS are two more Shopify stores that use Osano.

The first round covered Stores K and R on 2026-10-20 and 2026-10-21 (UTC+8), with a Fresh, Reject, and Accept run on each. The second round ran on 2026-11-15, between 17:16 and 17:30 PST, on all four stores. We reran all four the same evening, between 18:56 and 19:38 PST, over HTTP/1.1, with five runs per state. The table and delivery counts below come from that rerun. Every run went direct from the US with no proxy, in headless Chrome with a clean profile, en-US.

On the home page we waited for the banner, clicked about 4 seconds later, and waited 12 seconds. Then we opened one product page. In November we stayed on the product page about 91 seconds. Request bodies come from the page's request log, which kept the first 4,000 characters of each body. Whether a request reached the server comes from Chrome's network log.

In November only Store AP showed a banner to our US visitor. On Stores K, R, and AS, Osano started with all four categories set to Accept and sent one /record request as the page loaded. For Stores R and AS we rejected through the footer link that opens Osano's settings panel. We switched off Targeted Advertising, Personalization, and Analytics, switched on the opt-out toggle, and pressed Save. Store K's footer link opened a panel with no toggles, so we have no November Reject run for Store K.

November rerun: tattle attempts and what got through

Swipe to see every column

Tattle and /record requests, five runs per row, HTTP/1.1, November 15, 2026 (90 seconds on the product page)

StoreRunWhat we didtattle attemptstattle delivered/record sent (delivered)consented value
Store KFreshNo banner005 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS
Store KNo banner, no clickFive more loads005 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS
Store KSettings panelSettings panel, no switches, nothing saved005 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS
Store RFreshNo banner1095 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS
Store RNo banner, no clickFive more loads10105 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS
Store RRejectSettings panel: three categories off, opt-out on, Save10910 (10)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS; then ESSENTIAL, OPT_OUT
Store ASFreshNo banner10105 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS
Store ASNo banner, no clickFive more loads10105 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS
Store ASRejectSettings panel: three categories off, opt-out on, Save10910 (10)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS; then ESSENTIAL, OPT_OUT
Store APFreshBanner shown, no click550 (0)no /record
Store APRejectBanner Deny555 (5)ESSENTIAL, OPT_OUT
Store APAcceptBanner Accept555 (5)ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS

Stores R and AS each made two tattle attempts per run, one about 30 seconds after the home page opened and one on the product page, about 30 seconds in on most runs. Store AP made one per run, on the product page. Store K made none in 15 runs, including 91 seconds on each product page.

72 of the 75 attempts went through, with 204 on the preflight and 204 on the POST. After a report went through, the browser made no further tattle attempt on that page. The other 3 were on the product pages of Stores R and AS. Their preflight got no answer in the 60 seconds before we closed the browser, so Chrome never sent the POST. None of the 75 got a 403.

Every /record request we logged reached Osano with 204, including all 65 in the HTTP/1.1 rerun.

What the tattle body held

In October, each Store R entry had six fields: configId, customerId, currentURI, language, storeType, and storeKey. In November, entries on Stores R, AP, and AS also had a seventh, revision, and Store AS's cookie entries added the cookie's domain, path, expires, samesite, and secure settings, plus max-age on some. configId and customerId point to the site's Osano setup and account; the /record request carries the same two values under different names.

storeType was "script" on Store R, "localStorage" on Store AP, and mostly "cookie" with a few "script" entries on Store AS. For scripts, storeKey was the full script URL. For cookies and storage, it was the key name. currentURI was the page URL. No readable entry on any store held a consent value or the userConsentId sent to /record.

Store R's product page report was the same in every run we could compare. Its 13 readable entries named nine scripts from Okendo's CDN, Instagram's embed.js, TikTok's embed.js, one TikTok CDN file, and one script from the store's own domain. The first 4,000 characters matched across the October Fresh, Reject, and Accept runs. In the HTTP/1.1 rerun, all five product page reports after Reject had the same body, and 7 of the 10 runs without Reject matched it.

Where the Reject choice went

On Store AP, pressing Deny sent one POST to consent.api.osano.com/record with consented set to "ESSENTIAL, OPT_OUT". Accept sent "ESSENTIAL, MARKETING, PERSONALIZATION, ANALYTICS". On Stores R and AS, pressing Save in the settings panel sent a second /record request with "ESSENTIAL, OPT_OUT", after the one sent at page load. In October, Stores K and R sent /record only on the click, with the same two values.

The /record body had osnoCustomerId, osnoConfigId, userConsentId, extUsrData (empty), and consented. Store K's also had consentingDomain.

Why our first runs showed a 403

The October round and our first November run used Chrome over HTTP/2 on a test machine that resolved every hostname to the same address. In that setup Chrome sent requests for different hosts over one shared HTTP/2 connection, and CloudFront answered the tattle preflight with 403. That is why 38 of 39 attempts failed in the first November run. On the same machine the same evening, over HTTP/1.1, none of 75 attempts got a 403. We now count those 403s as an artifact of our setup.

Check your own Osano store

  1. Open DevTools, go to Network, turn on Preserve log, and filter by osano.com.
  2. Reject through the banner or through the link that opens Osano's settings. Find the POST to consent.api.osano.com/record and open its Payload. The consented field shows the categories Osano saved.
  3. Stay on a page for at least 40 seconds. A POST to tattle.api.osano.com may appear on its own timer. Check the OPTIONS row before it; a failed preflight means the POST was never sent.
  4. Open the tattle Payload. The storeKey values are what Osano reported. Compare them with the Discoveries list in your Osano dashboard.
  5. Under Application, Local Storage, look for keys that start with osano_consentmanager_tattles.

A storeKey entry after Reject means Osano noticed the script, cookie, or key. To see whether a script ran, check that vendor's own requests.

Limits

We tested four stores from a US connection on two days. The HTTP/1.1 rerun had five runs per state, and the earlier rounds had one. In November only Store AP showed a banner, Store K gave us no Reject run, and Stores K, R, and AS had no banner to click for Accept. We read the first 4,000 characters of each body. Three preflights in the rerun got no answer before we closed the browser, and we did not retry them. We did not check any store's Osano mode, test other regions, or see what Osano's servers do with a report.

FAQ

What is tattle.api.osano.com?

An Osano host. Osano's CSP guide lists it among the subdomains its Consent Manager uses, and Osano's release notes describe "tattles" as grouped reports sent on a schedule. On the stores we tested, each report entry named a script URL, a cookie name, or a storage key, plus the page where Osano saw it.

Does tattle.api.osano.com receive my Reject choice?

Not in any entry we could read on Stores R, AP, or AS. The entries had no consent field. The choice went to consent.api.osano.com/record, with the value ESSENTIAL, OPT_OUT after Reject.

Why does tattle.api.osano.com still appear after I click Reject?

On Stores R, AP, and AS, the browser sent a tattle report about 30 seconds after a page load, before and after Reject. On Store R, every product page report after Reject had the same body as most runs without Reject.

Why does the tattle request show a 403 or CORS error?

In our HTTP/1.1 rerun it did not. 72 of 75 tattle attempts got 204 on the preflight and 204 on the POST, and none got a 403. The 403s in our earlier runs came from our test setup, where Chrome sent several hosts over one HTTP/2 connection. If you see a 403, open the OPTIONS row and check whether the request went through a proxy or VPN.

Should I block tattle.api.osano.com in my CSP?

Osano's CSP guide lists it among the subdomains to allow. Osano's release notes tie tattles to the queue of uncategorized scripts the site owner classifies, so blocking it would stop those reports from visitors' browsers.

Sources

These links cover the platform and regulatory context used in this guide. Applicability still depends on the organization and jurisdiction.

What tattle.api.osano.com Sends After Reject | ConsentProbe