# Leak Tracing (/docs/leak-tracing) Found one of your PDFs shared where it shouldn't be? Upload the copy you found on the **Leak tracing** page, in Imprint's menu between **Orders** and **Upgrade**. Imprint looks for the codes and customer details your downloads carry and tells you whose download it was. ![Leak tracing page](/images/leak-tracing.webp) --- ## ๐Ÿ“ค Uploading a File [#-uploading-a-file] 1. Download the leaked file and keep it exactly as you found it. Don't print it or save it again. 2. In **Trace a leaked file**, drag the file onto the upload box or click it to choose the file. 3. The result page opens and updates on its own. The analysis usually takes less than a minute. - **PDF files only**, up to **500 MB**. Uploaded files don't count toward your storage. - You can analyze up to **3 files at a time**. Upload the next one when one of them has finished. - Found images or screenshots rather than a PDF? Save them as a single PDF first, with the pages showing the QR code among the first three. --- ## ๐Ÿ”Ž What Imprint Looks For [#-what-imprint-looks-for] Imprint reads the text of the first 100 pages and the file's document properties, looking for: - **Copy codes and licence codes** of your downloads: in a visible watermark, in the hidden [tracing marks](/docs/tracing-marks) and in the document properties. - **Your customers' email addresses and order numbers** (like `#1001`), for example in a visible watermark. - **The Imprint QR code:** Imprint also looks at the first 3 pages as images and tries to read the QR code. A QR code that's found counts as a high-confidence match, the same as a copy code in the text. Thanks to the QR code, copies that were printed to images, scanned, photographed or turned into screenshots can still be traced, as long as the QR code option was on when the copy was downloaded and the code is still visible and readable on those pages. Every clue is compared with **your own store's orders only**. Clues that don't match any of them are ignored. Your store's own email addresses (the ones Shopify has for your store, and the sender address of your own email server) are never clues. They often appear in your PDFs, for example as a support address, and the test orders you place with them are left out. --- ## โฑ๏ธ Statuses [#๏ธ-statuses] | Status | Meaning | | :--- | :--- | | **Waiting** | The file is in line for analysis. | | **Analysing** | Imprint is reading the file. | | **Done** | The analysis finished. Open it to see the result. | | **Failed** | The file couldn't be analyzed. The result page says why. | The most common reasons for **Failed**: the file isn't a readable PDF, it needs a password to open (remove the password and upload it again), or the analysis took too long and found no clue in the time it had (try again, or upload a shorter part of the file). If a long file takes too long to read in full but some clues were already found, the analysis still finishes as **Done**. The result then says that only part of the file's text was searched. The clues it shows come from the document properties, the QR codes and the text that was read in time. --- ## ๐Ÿงพ Reading the Result [#-reading-the-result] ![A leak lookup result](/images/leak-tracing-result.webp) When a match is found, **Whose copy this is** shows: - **The confidence** of the match and what it means (see below). - **Order:** the order number, when it was placed and the customer's name and email. - **Download:** the product and file, when it was downloaded, the country and device, and the **Copy Code**. If the download belongs to an order that is no longer in Imprint, the result still shows the download but can't show the customer. Below that you'll find: - **Some clues point to other orders:** the file also holds clues of other orders, for example when the customer bought more than once, or when copies were combined or shared between customers. Once a code has identified the order, the customer's other orders that only their email points to are left out. - **What was found:** every clue, where it was found (the document properties, or the page text and on how many pages) and its own confidence. - **Uploaded file:** the size, page count, and when it was uploaded and analyzed. ### No match [#no-match] **"No tracing code, customer email address or order number of your orders was found in this file."** The usual reasons: - The copy was downloaded before tracing marks were on. - Its watermark and document properties were removed. - It was printed to images or scanned and had no readable QR code, so there was no text or code left to read. Turn on [tracing marks](/docs/tracing-marks) and the QR code so future copies can be traced. --- ## ๐ŸŽฏ Confidence Levels [#-confidence-levels] | Level | What it means | What to do | | :--- | :--- | :--- | | **High confidence** | A copy code of this download is in the file (in the text, the document properties or the QR code), or the order's licence code is in the document properties. | You can act on it. | | **Medium confidence** | The order's licence code is in the page text, or the customer's email address is in the file, but no copy code. | Check the details before you act. | | **Low confidence** | Only weak clues point to this order. | Check the details before you act. | A licence code in the page text is only a medium clue, because an ordinary word or number in the text can match a licence code by chance. An email points to a customer, not to a single download. If they bought more than once, every order with that email is listed and Imprint can't tell which one the file came from. An order number on its own is a weak clue, so Imprint only counts it when another clue already points to the same order. Before you act on a **medium** or **low** match, open the file yourself, look at the watermark, and compare the product and dates with the order's **Download Log**. --- ## ๐Ÿ—‚๏ธ History and Order Pages [#๏ธ-history-and-order-pages] - **Recent lookups** lists your 50 most recent uploads with their **File**, **Status**, **Order**, **Confidence** and **Uploaded** date. - Lookups stay until you delete them. **Delete lookup** removes the lookup and its result. The order and its downloads are not affected. A lookup can be deleted once its analysis has finished, not while it's still **Waiting** or **Analysing**. - Each order page has a **Leak Lookups** section listing the leaked files traced to that order, with a **Draft a takedown notice** button. See [Managing Orders](/docs/orders). --- ## ๐Ÿ” Your Uploads Stay Private [#-your-uploads-stay-private] - The file is deleted as soon as it has been analyzed, whatever the outcome. Only what it matched in your orders is kept. - A lookup stores no customer details. The customer shown on the result is read from the order, so if their data was erased at their request, the result shows the erased details too. --- ## ๐Ÿšจ What to Do When You Find a Leak [#-what-to-do-when-you-find-a-leak] 1. **Save the evidence.** Note the web address of every page or download link that shares the file, and download a copy. 2. **Trace it.** Upload the file here and open the result. 3. **Check the match.** High confidence points to one download. For medium or low, check the details first. 4. **Stop further downloads.** Click **Restrict this order's downloads**. The customer's download links stop working and they see a message, which you can write yourself. You can also contact the customer. 5. **Ask for removal.** Click **Draft a takedown notice** and send it to the website. See [Takedown Notices](/docs/takedown-notices). 6. **Protect future copies.** Turn on [tracing marks](/docs/tracing-marks) and the QR code if they aren't on yet. --- ## ๐Ÿ”— Related Pages [#-related-pages] - [Copy Tracking](/docs/copy-tracking): copy codes, licence codes and Find Order by Code. - [Tracing Marks](/docs/tracing-marks): hidden codes and the QR code. - [Takedown Notices](/docs/takedown-notices): draft a DMCA or EU notice.