Privacy policy

Last updated November 10, 2026

InternScout is an internship search tool made by a UMass Amherst student. Searching is free and needs no sign-in; AI features have a free monthly allowance, and optional paid plans raise it. This page explains what data it uses, where that data goes, and how to delete it.

Searching listings

• You don't need an account or sign-in to search.
• The dashboard is a static page hosted on GitHub Pages. The listings are public job posts.
• Your saved statuses (like Interested or Applied), your filters and your profile are saved in your browser's local storage. Which listings you saved, and what you marked them, is never sent to us. If you arrive through a classmate's invite link, its invite code is kept there too, and is sent to us only when you sign in, to accept the invite. Delete my data clears it. Once the dashboard has suggested inviting a classmate, it also keeps the date, or that you waved it off or opened your link, so the suggestion appears only once; that is never sent to us. If you hide the dashboard's suggestion to review the extension, it keeps that one setting so the suggestion stays hidden; that is never sent to us either. Once the dashboard has shown its one suggestion to get the weekly email, it keeps that it was shown, dismissed or opened, so it appears only once; that is never sent to us. After you open a posting, the site may suggest the Auto-Apply extension once per browser session; your browser keeps that it was shown for the session, and the date if you dismiss it, so it stays away for 14 days. Neither is sent to us. If you press that suggestion's add-to-Chrome link, or a page's "Auto-Apply this" link, your browser also keeps which posting it was (its link, company and title) so the extension can offer to Auto-Apply it once it is set up. The extension removes it when it reads it. If it hasn't after three days, the site deletes it the next time you open the dashboard or one of its internship pages. It is not sent to us. If the extension is installed, the dashboard also hands it your profile, and the extension saves that with your Deep Dive (see below). If you are signed in on the dashboard but not in the extension, the dashboard also offers the extension that sign-in; the extension checks it with our server and uses it only if you press "Use my dashboard sign-in" in the extension. Pressing it gets the extension its own sign-in for the same account from our server, so signing out of the dashboard or the extension leaves the other signed in. If you hide paid listings, or hide the notice that they started, your browser keeps that one setting so they stay hidden; neither is sent to us.
• When you sign in, the dashboard keeps your sign-in in local storage too: a random sign-in code, a short account ID, and your email address, only so it can show which account is signed in. It stays until you sign out, press Delete my data, or the sign-in ends.
• One thing does go: the list of states you picked. While you are signed in, the dashboard sends just those two-letter state codes (plus REMOTE, if you ticked it) to our Worker, stored against your scrambled sign-in ID. It is sent once per change, not per listing. The server accepts nothing else on that route, and the scanner only ever reads the totals per state, so it can fetch fuller listings where students are actually looking.
• GitHub hosts the page and may keep normal web server logs under its own policy. The dashboard's fonts come from our own site; it loads its code library (React) from a public server, jsDelivr.
• We count page visits with Cloudflare Web Analytics. It uses no cookies and doesn't identify you: it sees which page was loaded, the referring site, your browser type, a rough country and how fast the page loaded.
• Because Cloudflare only samples visits, our own server also counts them. Each page load sends the page's address, the name of the site that linked to it (such as google.com, not the full link) and any short campaign tags on the link. We keep only daily totals by kind of page and where visits came from, for example "landing pages, from search: 12". If a tagged link brought you here, its short source, medium, campaign and content labels stay in this tab until you close it. They have no visitor ID and are used only to add anonymous daily totals for a Chrome Web Store click, a weekly email form submission or an employer order form submission. The detailed campaign rows are deleted after 400 days. Those totals do not say that an extension was installed, an email subscription was confirmed or an order was paid. The same way, we count how many times a day a few steps happen (pressing a link to the Chrome Web Store, pressing a link to review the extension there, opening a posting, seeing or pressing the "Add to Chrome" suggestion that can appear after you open one, seeing a weekly email invitation, pressing its link or sending its sign-up form (only the invitation's position and that it was seen, pressed or sent, never the address you typed or whether you confirmed), our server seeing an account for the first time, signing in on the site or in the extension, saving a first profile, starting Auto-Apply, a first Auto-Apply of the month, reaching a monthly limit, opening your invite link, opening a plan's checkout, buying a plan), as totals that say nothing about who.

  When these features and their counters are enabled, we also count expanding a company's results, choosing a major, saving a role, opening My list, pressing a share link, opening our GitHub list or widget page, seeing the out-of-runs panel, and pressing its invite, top-up or plan controls. This includes opening or sending a company or search email-follow form, reaching its confirmation page, playing the demo, and pressing Print on a flyer. Each of these browser steps is counted at most once per page load, as a daily total by step and kind of page. These stored totals contain no account ID, email address, company, role, chosen major or search terms. They do not prove that an email subscription was confirmed, a purchase completed, or a flyer printed. These aggregate totals have no automatic expiry and cannot identify a person's activity for individual deletion.

  We count paid listings the same way: how many times a day each one was shown and opened, by kind of page, and nothing else. Your IP address is seen only to stop floods of fake visits; it's held in memory for a minute and never saved. If you open any page with ?nocount=1 at the end of its address, your browser keeps one setting that turns this counting off for it; ?nocount=0 turns it back on.
• We don't use ad trackers, and we set no cookies of our own.

Your profile and files

• Your resume, CV, transcript, cover letter and writing samples are stored on your device only. They live in browser storage (dashboard) or Chrome extension storage. We keep no copy of them, and they are never uploaded for safekeeping or backup.
• Your profile (majors, class year, work authorization, the facts and stories from your Deep Dive and your saved answers) is kept on your device too, and by default the extension also saves a copy to your InternScout account. What is saved and what never is: Your Deep Dive, saved to your account.
• Some of these files are sent to the AI when you run a feature that reads them. They go inside the request, through our Worker to Google's Gemini API, and are not stored at either end (see AI features). Which file goes depends on the feature and on the file's format:
  • The Deep Dive reads your resume and CV to fill in your profile, and your cover letter and writing samples to learn your writing voice. A PDF is sent as the file itself. A Word or plain-text file is read on your device first and only the text it contains is sent, up to 60,000 characters.
  • Tailoring your resume sends your resume when it is a PDF: the whole PDF goes, so the AI can transcribe its layout and the tailored copy can keep it. That reading is cached on your device and reused, so the file is sent once per resume, not once per application. If your resume is a Word file, the extension opens it locally and sends only the bullet text; that file never leaves your device.
  • Your transcript is never sent to the AI. It is only ever attached to an employer's own form when their site asks for one.
• The extension can create accounts on employer application sites for you. It saves the sign-up email and those site passwords in extension storage on your device. The passwords are never sent to us or to the AI: the extension types them into the page itself, and the AI is only told to ask for one, never given it. The sign-up email address is included in each Auto-Apply request, because the AI has to know which address to register or sign in with.
• The extension keeps a count of the applications it has filled on this device, and what you chose on its suggestion to leave a Chrome Web Store review (Leave a review, Not now or Don't ask again), so it asks at most twice. This stays in extension storage and is never sent to us. The only thing it tells the dashboard, inside your browser, is whether the dashboard may show its own review suggestion.
• Chrome does not encrypt extension storage. Anyone who can use your computer account could read it.
• We have no copy of your files, your demographic answers or your saved logins. If you clear them, they're gone. Your profile can be restored from your account if you left saving on.

Your Deep Dive, saved to your account

When you are signed in, the extension saves the profile your Deep Dive builds to your InternScout account, so it's there when you sign in on another computer or reinstall the extension. This is on by default.

• What is saved: the profile the Deep Dive builds. That is the facts it collects about you (such as your name, contact details, address, work authorization and the other standard application questions), your education, experience, projects, skills and links, your stories and goals, your writing voice, and your answers to its questions. It also includes the profile you set on the dashboard (your majors, class year, graduation term, the kinds of roles, terms and states you picked, and your work authorization), and whether and when you finished the Deep Dive.
• What is never saved to your account: your resume, CV, transcript, cover letter, writing samples and any other file; your answers to the six demographic questions (gender, race, Hispanic or Latino, veteran status, disability and LGBTQ+); the logins and passwords the extension saved for employer sites; your own API keys; and your list of past application answers and jobs. These stay on your device. Our server also removes the demographic answers from anything it is sent, as a second check.
• Encrypted at rest. The saved profile is encrypted on our server with a key made for your account alone. It is stored against your scrambled sign-in ID, with the time it was last changed and its size. It is decrypted only to send it back to your own extension while you are signed in.
• It is not used for anything else: not for the AI (the extension sends the AI what each feature needs from the copy on your device, as described under AI features), not for ads, and not shared or sold.
• Turning it off: untick Save my Deep Dive to my account in the extension. It asks you to confirm, then deletes the copy on our server right away and stops saving. Your profile on the device is untouched.
• Delete my data (in the extension or on the dashboard) deletes the saved profile too, along with every sign-in, so every device is signed out.

AI features

Tailored resumes, the Deep Dive, Auto-Apply and short answers use AI. Here is what happens when you run one:

1. The extension sends what the task needs to our server, a Cloudflare Worker. That is more than the questions alone, and it is not the same for every feature, so here is what each one sends. Two of them send your whole stored profile (your name, contact details, education, experience and the stories from your Deep Dive): Auto-Apply, which adds your writing voice, and the Deep Dive's interview step, which leaves out your voice and your saved answers. The rest each send only what their own bullet names below, which for one tailoring path is still most of your experience and education:
   • Auto-Apply sends your whole profile on every turn, and the sign-up email it uses for that site. On the first turn it also sends the posting: company, title, location, apply link and the first 5,000 characters of the description. Then, on each step, a snapshot of the application page: its web address and page title, the address of any form frame inside it, the headings, which step of the form you are on, whether a CAPTCHA is showing, any error messages the site is displaying, and every visible form field with its label, the question around it, what is currently in it, its options, whether it is required, its placeholder text, its length limit and its own error. It also sends the buttons on the page and a list of anything the extension has already filled in for you. When a page has fewer than four fields (a sign-in wall, an interstitial, a page that failed to load), it sends up to 2,500 characters of the page's visible text instead, so the AI can tell what it is looking at.
   • Tailored resumes send the job posting's title, company and the first 6,000 characters of its description, your career goal and your writing voice, and your resume: the whole PDF file if it is a PDF, or just its bullet text if it is a Word file. Your stored profile is not sent on this path, but the resume file is: a PDF goes whole, so whatever is printed on it (your name, your contact details, your education) goes with it, and a Word file is read here on your device and only its bullet text is sent. If you have no resume on file, the fixed template is built instead from your experience, projects, skills, education, career goal and voice; your name and contact details are printed onto it here on your device and never leave it.
   • The Deep Dive sends something different at each step. Read my resume sends your resume and CV files and nothing else; on a first run there is no profile yet, because that call is what fills it in. The interview sends your profile (all of it except your writing voice and your saved answers) along with the conversation so far. Analyze my writing sends your writing samples, your cover letter and up to 20,000 characters of your own interview replies. Finish & save stories sends the interview transcript plus the stories and goals you already have.
2. The Worker checks your sign-in and your monthly allowance.
3. It forwards the request to Google's paid Gemini API and passes the reply back to you.

Auto-Apply only reads a page while it is working on an application you started, in the tab it opened for that job. It does not watch your other tabs, and nothing is sent for a page you merely have open.

• We don't store or log what you send or what the AI replies. The Worker only records that a task ran and how many tokens it used, so we can track cost.
• No training. We use Google's paid API. Under Google's terms for paid services, Google does not use your prompts or replies to improve its products. Google may keep them for a limited time to detect abuse, under its own terms.
• Cloudflare runs the Worker and handles the network connection under its own privacy policy.

Using your own key (advanced). If you enter your own Anthropic or Google API key in the extension, requests go straight from your device to that company. They don't pass through our server, and that company's terms apply.

Sign-in

• You sign in only to use AI features, to save your chosen states and to save your Deep Dive to your account. Search never needs it.
• You sign in with any Google or Microsoft account. You type your password on the provider's own page. We never see or receive your password.
• The provider gives back a short-lived token. The Worker checks it and does not save it. It reads the email in the token only to see whether it is a verified school (.edu) address, which gets a larger AI allowance. The email is never stored on our server.
• Staying signed in. The Worker then gives your browser or extension a random sign-in code, and that code is what signs you in from then on, so you stay signed in on that device until you sign out. Your device keeps the code, with your email address so it can show which account is signed in. Our server keeps only a fingerprint of the code (a one-way hash it can't be turned back into), your scrambled sign-in ID, your allowance tier, whether you used Google or Microsoft, and when the sign-in was made, last used and ends.
• A sign-in ends when you sign out, when you press Delete my data (which ends all of them), or after a year without being used. Each device or browser has its own, and we keep at most 20 per account, dropping the oldest.
• If our server can't start a sign-in like this, the dashboard falls back to keeping the provider's token for the current browser session only. It's gone when you close the browser or when it expires, about an hour later.
• From the token, the Worker makes a one-way hash: a scrambled code that can't be turned back into your name or email. We don't store your name or email.

What we store

This table covers your student account. If you have joined a paid plan, add the plan and billing records listed under If you join a paid plan. Employer interest forms and paid-listing orders have separate records described below.

Separately from the table, we keep running totals of what AI costs the whole service: one for the month across the whole service, and one for the day's free-allowance share, so a single busy day can't empty the month's budget. When a month's per-run counts are deleted, we keep how many runs of each AI task that month took 1, 2, 3 to 4 and so on AI calls, counted across the whole service. None of these is linked to any account, and none says anything about who used what. The same goes for the daily totals of how often each paid listing was shown and opened (see Paid listings).

Apart from what is inside your saved Deep Dive (which has your name and contact details in it, encrypted), and what an employer sends us through the employer interest form or a paid-listing order, we don't store names, emails, passwords, resumes, files, prompts or AI replies. We don't sell or share data. The only paid content on InternScout is the labelled paid listing described under Paid listings, and it is never chosen from data about you.

If you sign up for the weekly email, Buttondown keeps your address, not our server. See If you sign up for the weekly email.

If you join a paid plan

The paid plans are optional. Supporter is $4/month and Pro is $8/month; both raise your monthly AI allowance and nothing else. Search and the free AI allowance never need one.

• Stripe handles the payment, not us. Pressing Supporter or Pro sends you to Stripe's own checkout page. Your card number, billing name and billing email go to Stripe and are never sent to, or seen by, our server. Changing or cancelling happens on Stripe's own billing page too.
• All Stripe learns about you from us is your hashed ID, so a payment can be matched to an account without Stripe being told which student it is.
• Stripe is the seller of record for the paid plans, so its checkout also asks for your name and billing address to work out sales tax or VAT. That stays with Stripe; we are never sent it.
• What we keep: your hashed ID, which plan you are on and whether it is active, the Stripe customer and subscription IDs, the date it is paid through, and the IDs of payment events we've already processed (so a repeated message can't charge or credit you twice).
• Stripe processes payments under its own privacy policy.
• While a subscription is live, Delete my data is refused with a message asking you to cancel first, so nothing keeps billing a card for an account that no longer exists here. Cancel on Stripe's billing page, then delete.

If you sign up for the weekly email

The weekly email is optional. Search, your profile and everything else on InternScout work without it. You sign up with a form: under the listings on the dashboard, on each field's page and the "New internships this week" page under internscout.org/internships, or on internscout.org/digest. Every one of them works the same way, as described here.

• What it contains: the internships InternScout found in the past week, grouped by field: the fields with the most new roles that week, each with a few postings and a link to the rest. For now everyone gets the same email. The fields you tick are saved with your subscription so it can be matched to them later.
• How often: once a week. Buttondown first sends one confirmation email, and nothing else arrives until you click its link.
• Buttondown holds your address, not us. The form sends your email address, and the fields you ticked, straight from your browser to Buttondown, the service that sends the email. They never pass through or get stored on InternScout's servers, which is why the What we store table above has no row for them. Buttondown keeps them under its own privacy policy.
• The person running InternScout can see the subscriber list in Buttondown's dashboard: your address, the fields you ticked, whether you have confirmed, and what Buttondown records when you sign up: the IP address you signed up from, a rough location worked out from it, and the page the form was on. Open and click tracking is switched off for this email. The list is used only to send this email, is never sold or shared, and is not copied anywhere else.
• Past emails are public. Each week's email is also posted on the newsletter's web archive at Buttondown, which the sign-up forms link as "See last Monday's email", so you can read one before signing up. The email is the same for everyone and has nothing about any subscriber in it. The archive is Buttondown's page, so opening it is a visit to Buttondown under its privacy policy.
• Signing up again with other fields adds them to the ones you had; it doesn't remove any. To drop a field, write to us, or unsubscribe and sign up again.
• Unsubscribing: every email has an unsubscribe link at the bottom, and it stops the emails. To have your address deleted from Buttondown's list as well, write to us at the address under Contact below.

If you fill in the employer interest form

The form on internscout.org/employers is for companies that want to ask about a paid listing, or to have their job board added or their details corrected, which is free. Students never see it or need it. Sending it costs nothing and commits no one to anything.

• What we keep: what you type into the form (your name, work email and company, and, if you fill them in, your role or title, who you're hiring for, the boxes you tick and your message), and when you sent it. It is stored on InternScout's server (a Cloudflare database). Your IP address is not kept with it.
• Why: only so the person running InternScout can read it and contact you about InternScout for employers. It isn't added to the weekly email, sold, shared, or used for anything else.
• Stopping spam: to stop one place from flooding the form, our server counts how many times it was sent from each network address that day. The count is kept under a scrambled form of the address (a salted hash), never the address itself, and is deleted within two days.
• How long: up to a year from when you sent it, then it is deleted automatically.
• Deleting it sooner: write to us at the address under Contact below, from the email you used or naming it, and we'll delete it.

If you buy a paid listing on our site

Students never see or need this. When an employer orders a paid listing on internscout.org/employers/buy:

• What we keep: your name, work email and company, the roles and dates you chose, whether you said you are a nonprofit or public body, any pilot code, which version of our terms you accepted and when, and the Stripe references for your payment and any refund. Stripe handles the card and the invoice; card details never reach us. Our server keeps the link we email you only in a scrambled form (a hash).
• Why: to confirm the email is yours, to review and run the paid listing, to send you emails about it (confirmation, approval or refusal, a completed campaign's aggregate report, and the choice of credit or refund if the role closes early), and to keep a record of the sale. A campaign report may link to the order page if you choose to promote another open role; it never renews or charges you automatically.
• Report emails: an unsent campaign report's address and message are removed after 30 days. Once sent, the queued copy is erased; a small record of when it was sent stays with the order. Counts are rounded totals, not information about individual students.
• How long: an order whose email is never confirmed is deleted after 7 days. A confirmed order is kept for 7 years as the record of a sale, then deleted.
• Who else sees it: Stripe, which takes the payment, and Cloudflare, which runs our server and sends our emails. Stripe also handles your payment details under its own privacy policy (stripe.com/privacy). It isn't added to the weekly email, sold or shared with anyone else.
• Deleting it sooner: we have to keep the record of a sale, but you can ask us to remove your name and email from it once the paid listing has ended; we keep the company, amounts and dates.
• Stopping spam: orders are counted per network address per day the same way as the interest form above.

Paid listings

Some internship pages, and the dashboard after you search or pick a field, can show one box headed "Paid listing", with "Ad" before the company name. It means the employer has a business arrangement with InternScout to show that role in that spot: it paid, or it is trying paid listings in a free pilot.

• Only roles already listed: the box always points at a role InternScout already found on the employer's own job site, and the same role is also in the page's normal list. Paying can't add a listing or change one.
• Nothing else changes: paying never changes match scores, the order of the list, filters, counts or what Auto-Apply does.
• At most one per page, and it goes away on its end date. A box is dropped if the role closes, its deadline passes, its link changes, or it says it is unpaid (unless the employer is a nonprofit or government body).
• Nothing about you goes to the employer: "Open posting" goes straight to the employer's own posting, like any other link on the page. We count how many times the box is shown and opened, never who saw or opened it (see below).
• When it shows on the dashboard: only after you search for something or pick a field, and only for a role that is already in your results (when you search, among the 50 closest matches). If more than one could show, each has the same chance; the price never decides.
• It isn't chosen from your data: which paid listing you see depends only on what you searched for or picked and the page you're on. It is never chosen using your majors, class year, saved or applied listings, Deep Dive, sign-in, plan, or anything in the extension. The extension never shows paid listings.
• Your saved profile can only hide one. If you have saved a profile on the dashboard, the dashboard uses it, inside your browser, to hide a paid listing that doesn't fit you (your class year, citizenship, eligibility or season). It is never used to pick one, and it never leaves your browser.
• You can hide them: press "Hide paid listings" in the dashboard's box, or turn off "Show paid listings" under More filters. Your browser keeps that one setting, and it is never sent to us.
• What we count: how many times a day each paid listing was shown (at least half of its box on your screen, once per page load) and opened (its "Open posting" link pressed, once per page load), by kind of page: the dashboard or our internship pages. Your browser sends only the paid listing's id and the address of the page. These are totals that say nothing about who. Opening a page with ?nocount=1 turns this off too.
• What the employer gets: those totals, for a whole calendar month once it has ended, rounded down to the nearest 10. Any number under 10 is shown as "fewer than 10", and the totals are split by kind of page only when every number in the split is 10 or more. Employers never get anything about any student. We don't see applications, so we can't tell them who applied or how many did.
• What we keep about a paid listing: which listing it is, the employer's name, its dates, whether it is on, and the totals above. Nothing about any student. Daily totals are kept for 400 days, then combined into one total per month.
• No ad network, tracking pixel or cookie is involved. A paid listing is an ad, not a recommendation: check it the same way as any other posting.

Delete your data

1. On our server: sign in, then press Delete my data in the extension or on the dashboard. Your chosen states, your saved Deep Dive, every sign-in (so every device is signed out), your plan record, your invite code and extra runs, and everything from earlier months are removed right away. Three invite records stay, each just your hashed ID with a date or a number, because without them deleting and signing in again would get around the invite rules: the date we first saw your account (so it can't count as new again); if you joined through an invite, the record that you did, without the inviter's ID (so you can't accept a second one); and if your invites earned you extra runs, how many did (so the limit of 10 can't start over). This month's counts (your hashed ID and a few numbers, nothing else) stay until the month ends, because removing them on request would let anyone reset their limits by deleting and signing in again. They are then deleted automatically, within a day of the month ending.
2. In your browser: clear site data for internscout.org (and for bpmcginley.github.io, the address InternScout used before September 2026, if you visited it then). In Chrome, click the icon left of the address bar, then Site settings, then Delete data. This removes your dashboard profile and saved statuses.
3. In the extension: remove it at chrome://extensions. Chrome deletes its storage, including your files and saved site passwords. That doesn't reach the copy of your Deep Dive saved to your account: turn saving off or press Delete my data first.
4. The employer interest form: write to us at the address below and we'll delete what you sent. Delete my data doesn't cover it, since it isn't tied to a sign-in. It is deleted automatically a year after you sent it.
5. The weekly email: use the unsubscribe link in any of the emails. Delete my data can't reach it, because the address is kept by Buttondown, not by our server. To have it deleted from Buttondown's list as well, write to us at the address below.

We can only find your server rows through your sign-in, because we don't know who you are. If you can no longer sign in, your state picks stop counting after 90 days of no use and your sign-ins end after a year. A saved Deep Dive can only be reached through a sign-in to the same account, so delete it (or turn saving off) while you still can.

Chrome Web Store: limited use

The extension's use of data follows the Chrome Web Store User Data Policy, including the Limited Use requirements. In plain words:

• Data is used only for the extension's single purpose: helping you fill out internship applications, plus the AI help that goes with it.
• Data is only sent where that purpose needs it (to Google's Gemini API through our Worker, and your Deep Dive profile to your own InternScout account unless you turn that off), or where the law requires.
• Data is never sold.
• Data is never used for ads, including personalized ads. The extension never shows paid listings, and nothing in it is used to choose the ones the website shows.
• Data is never used to decide credit, lending or anything like it.
• No person reads your data, unless you give clear permission (for example, you paste it into a bug report), it's needed for security, or the law requires it.

Children

InternScout is built for college students. It isn't meant for anyone under 13. If you're under 18, a parent or guardian has to agree to the terms with you before you start a paid plan. Paid listings are never aimed at anyone, of any age: they depend only on what is searched for or picked and the page being viewed. If you believe a child under 13 has given us data, write to us at the address below and we'll delete it.

Changes

If this policy changes, we'll update the date at the top. Big changes will also get a notice on the dashboard.

Contact

InternScout is run by Bruce McGinley, an independent student project in Massachusetts, USA. Email brucepmcginley@gmail.com.

Questions, privacy requests or anything else: use the InternScout feedback form. It needs no account and only the person running InternScout reads it. Expect a reply within a few days; this is a one-student project, not a support desk.

For bugs you don't mind discussing in the open, you can also open a GitHub issue. Issues are public, so please don't put personal details in them.