Privacy Policy

GoSmart PDF has no server that receives your documents. Every file is opened, changed and saved by your own browser. Nothing is uploaded to a server, nothing is stored, and no account is needed. This policy explains how Mrityunjay Kumar, trading as GoSmart PDF, in India, handles the little information the site does receive.

Last updated: 5 September 2026

This Privacy Policy explains how Mrityunjay Kumar, trading as GoSmart PDF, in India, handles information in connection with the gosmartpdf.com website.

Our privacy-first design

GoSmart PDF is built to keep your documents on your own device. Every tool runs in your browser. The PDFs you choose are read into that browser tab, worked on there, and handed back to you as a download. They are not uploaded to us or to anyone else, and we hold no copy of them at any point.

There are no accounts. Nothing on this site asks you to register, and we hold no profile of you, no history of the tools you have used and no record of the files you have opened. When you close or reload the tab, the file is gone from the site entirely. There is nothing to delete and no retention period, because there was never a stored copy.

Information we process

Locally stored data. Your licence key and your choice of settings are stored in your browser on your own device. We do not receive them.

A device identifier. If you subscribe, your browser creates a random identifier and sends it with each licence check. It is not tied to you and says nothing about you. It exists only so one subscription runs on one device at a time, and so you can move it to another device yourself.

Payment information. When you subscribe, payment is processed by Dodo Payments, our merchant of record. We do not collect or store your card details. Dodo Payments shares with us limited transaction data, such as a payment status and the email address on the subscription, so we can confirm it. See the Dodo Payments privacy policy for how they handle your data.

Your licence email. When you subscribe we send your licence key to the address on the subscription, so you have a copy of it that is not on one device. The message is delivered by Resend, our email provider, who receives your address and the key for that purpose alone. It contains nothing about your documents, because we have nothing about your documents.

The AI model. Enabling the AI tools downloads a language model to your browser once, from a public content delivery network. That request brings the model to you and carries nothing about you or your documents.

The deeper redaction scan. Redact PDF can read each page with a model instead of only matching shapes, to find names, addresses, reference numbers and other personal details that no pattern can see. There are two, and neither is downloaded until you ask for it: a small one trained for this job, which comes from our own storage rather than a third party, and the larger AI model for details that have no shape at all. That request brings the model to you and carries nothing about you or your documents. Once the small one is on your device it is used by default, because a shapes-only pass over a document full of people finds nobody and then reports itself checked; you can switch it off on the tool, and it stays off. Both only ever add to what the patterns already found, neither can mark anything as safe, and the reading happens on your device. No page is sent anywhere.

One cookie, for that smaller model. It is our own rather than a third party one, so it is served from this site and only to a device that holds a subscription. Checking your licence key sets a cookie named gsp_model carrying that key and the device identifier above, and the browser sends it only on requests for the model files. Nothing else on the site reads it, it measures nothing and it identifies you to nobody. Without a subscription none is set, and clearing your site data removes it.

The reader. Asking a question across several documents downloads a second, much smaller model to your browser once, from a public content delivery network, in the same way as the language model. It turns each passage of your documents into numbers so a question can be matched to the passage that answers it. That happens on your device, and neither the passages nor the numbers are sent anywhere.

The voice. Having a document read out loud downloads a speech voice to your browser once, from a public content delivery network, in the same way as the language model. The document is turned into speech on your device and the audio file is made there too, so neither the document nor the recording of it is sent anywhere. The audio exists only in the page until you save it, and closing the tab discards it.

Your recordings. Asking a question across audio files downloads the same speech recogniser used for the microphone, once, from a public content delivery network. Your recordings are decoded and transcribed in your browser, the transcript is held in the page for as long as the tab is open, and neither the audio nor the transcript is sent anywhere. Nothing is written to disk and nothing survives closing the tab.

The microphone. Speaking an instruction is optional and off until you ask for it. Enabling it downloads a speech recogniser to your browser once, from a public content delivery network, in the same way as the language model. The recognition itself runs on your device: the microphone opens only while you are recording, the audio is turned into text in your browser, and neither the audio nor the text is sent anywhere.

Your saved workflows and run history. The Assistant keeps two lists in this browser so they are still there on your next visit: the workflows you choose to save, and a record of what you have run. The run record holds the step names, how many files went in and out, how long it took, and the names of any documents that could not be read. That last one is the only place this site keeps anything about a document of yours after the tab closes, and it is a file name and nothing more: no document, no page and no text is ever stored. It is capped at the fifty most recent runs, it is held on your device alone, it is never sent to us or to anybody else, and a Forget button on the same panel deletes all of it.

Files you make while you are here. A finished file is offered to the next tool you open, so that splitting a bundle and then signing the piece you wanted does not send you back through your downloads folder. Those files are held in the memory of the browser tab you are looking at, for as long as it is open, and nowhere else. Closing or reloading the tab ends them, nothing is written to disk by us, and nothing is uploaded. A Clear button on the row removes them sooner.

A file shared to this app from another app. On a phone where GoSmart PDF has been added to the home screen, the system can hand it a file you picked in a chat or a file manager. The browser delivers that as a post, and this site intercepts it on your device rather than letting it reach a server: the file is written to the browser's own storage, on your phone, and is taken back out the moment the page reads it. It is never uploaded, and nothing about it is sent to us. An interrupted share leaves a file in that storage until the next one replaces it or you clear the site data.

Basic analytics and logs. Our host processes standard technical data such as IP address and browser type to serve and secure the site. Visitor numbers are counted by Cloudflare Web Analytics, which is cookieless and records no identifier that would let one visit be tied to another or to a person. Page views are also counted by Umami, which is cookieless and aggregate: the script that does it is served from this site rather than from anybody else, and it reports to this site, which passes the count on. Only the page path is recorded, never the rest of the address, because this app puts what you type into the address itself and that is yours. Your IP address is shortened to its network before it is passed on, so two readings can be told apart without either naming you, and nothing we know about you, no licence key, no device identifier, is attached to any of it. The only cookie this site sets is the model cookie described above. There are no advertising or tracking cookies, and nothing counts what you do inside a tool.

Where a subscription came from. When you subscribe, we record which one of our own pages your visit started on, as a single path such as /redact-pdf, and keep it on that subscription record. It tells us which pages are worth writing, which is the only reason it exists. It is written once and never changed, it is kept for nobody who does not subscribe, and until you subscribe it lives in that browser tab alone and is gone when you close it. A page address carrying anything you typed is refused rather than trimmed.

How we use information

We use the limited data we receive to operate the site, to confirm and support a subscription, to answer an email you send us, and to comply with legal obligations. We do not sell your personal data, and we do not use it to profile or advertise to you.

Sharing

We share information only with the providers who help us run the site: Dodo Payments, our merchant of record, for payments, Resend for sending your licence email, and Cloudflare for hosting and analytics. We also disclose information where the law requires it.

Data retention

Documents are never retained, because they are never received. Email addresses left on the notify form before it was withdrawn are deleted after two years. Payment and licence records are retained by us and by Dodo Payments as long as accounting and tax law requires. Data your browser stores stays on your device until you clear it.

Your rights

Depending on where you live, you have rights to access, correct or delete personal data we hold about you. To make a request, contact us at support@gosmartpdf.com and we will answer within thirty days. You can clear locally stored data yourself through your browser settings at any time.

Children

The site is not directed to children under 13, and we do not knowingly collect their personal data.

Changes

We may update this policy, and the last updated date at the top of this page will change accordingly.

Contact

Questions about this policy: email support@gosmartpdf.com, or see our Contact page.

Support and billing: support@gosmartpdf.com. Operated by Mrityunjay Kumar, India. Payments by Dodo Payments, our merchant of record.

Loading the editor…