NavBharatAI · Last updated 5 October 2026

Privacy Policy

Last updated: 5 October 2026

This Privacy Policy explains how NavBharatAI ("NavBharatAI", "we", "us", "our") collects, uses, stores, shares and protects your information when you use the NavBharatAI website (navbharatai.com), the NavBharatAI mobile applications, and every product surface inside them — including NavBharatAI Free chat, NavBharatAI Pro (the AI app builder), the Professionals assistants, the Nav App Store, and the developer tools (together, the "Platform").

We have written this policy to be read by real people, not only by lawyers. Where a section has legal weight, we have still tried to say it plainly.

By creating an account or using the Platform, you agree to the practices described here. If you do not agree, please do not use the Platform.

1. Who we are, and how to reach us

NavBharatAI is an India-first AI platform. For every question, request or complaint about your data, contact:

Email: info@navbharatai.com Subject line suggestion: "Privacy request — [your account email]"

This mailbox also serves as our Grievance contact for the purposes of Indian law, including the Information Technology Act, 2000 and the Digital Personal Data Protection Act, 2023 ("DPDP Act"). We aim to acknowledge privacy requests within 72 hours and resolve them within 30 days.

2. What we collect, and where it comes from

2.1 Information you give us directly

• Account information. When you sign in with Google, Apple, or a mobile number (OTP), we receive your name, email address and/or phone number, and a profile picture if your sign-in provider supplies one. We do not see or store your Google or Apple password. Phone-number login is verified by a one-time password; we store the number, never the OTP.
• Chats and prompts. The messages you type or dictate into any NavBharatAI chat (Free chat, NavBharatAI Pro build requests, Professionals assistants), including follow-up instructions, feedback and corrections.
• Files you upload. Documents, images, spreadsheets, PDFs and archives you attach — for example a .zip of an app you import into NavBharatAI Pro, a logo you add to a build, or a document you ask an assistant to analyse.
• Health-related documents and descriptions (web only). The Doctor AI / Senior Doctor Assistant surface (available on the website, intended for qualified medical professionals) accepts clinical descriptions and documents such as lab reports. This is sensitive information and is treated under the stricter rules in Section 5.
• Payment identifiers. When you buy tokens or a plan, the payment itself is processed by our payment provider (Section 7). We store the order identifiers, the amount, and the outcome — we never see or store your full card number, UPI PIN or banking password.
• Things you publish. If you publish an app to the Nav App Store, the app package, its listing (name, description, icon) and your developer name become part of the store listing you asked us to create. An app that runs in the browser (App Mart's "Play instantly") is shown with the display name on your NavBharatAI account, a short public creator code and the date it was published. The creator code is derived from your account but cannot be turned back into it, and your email address is never shown.
• App Mart likes, dislikes and comments. When you are signed in you can press 👍 or 👎 on an App Mart app and write comments and replies. Your comments are public: anyone who opens that app sees the comment with your display name, your profile picture (if you have one) and your creator code, and tapping your name opens your public App Mart profile — your name, picture, creator code and the apps you have published to App Mart. Everyone sees how many likes and dislikes an app has; only the app's creator sees who liked it, and nobody — not the creator, not other users — ever sees who disliked it. The creator is notified that you liked or commented. Your email address and account id are never shown on any of these screens. You can delete your own comment at any time, and you can block a person so their comments are hidden from you and they can no longer notify you. A comment containing abusive words is not posted; it is checked automatically against a word list when you press Post, and nothing is stored when it is refused.
• Your profile photo, name and bio. You can upload a profile photo and edit your display name and bio in Settings → Profile or on your App Mart profile. The photo, name and bio are public on App Mart — on your profile and next to your comments. Before an uploaded photo is saved it is checked automatically by an AI vision model for nudity, sexual content, graphic violence and hate symbols; a photo that fails, or that cannot be checked at that moment, is not saved. The photo is stored on our servers and deleted when you remove it or delete your account. A name or bio containing abusive words is not saved. A photo address pasted in Settings before uploads existed is shown only to you.
• App Mart follows. When you are signed in you can follow an App Mart creator. Everyone sees how many followers a creator has; only the creator sees who follows them, and the list of creators you follow is visible only to you. The creator is notified that you followed them, and you are notified (in the app, not by phone notification) when a creator you follow publishes a new app to App Mart. The "Liked apps" view shows the apps you pressed 👍 on — it is visible only to you. Blocking a person ends any follow between you in both directions.
• Support communication. Emails you send to our support address.
• Problem reports you send from inside the app. When you use "Report a problem", we receive what you typed, the kind of problem you picked, and any screenshot you chose to attach. So that the fault can be found without a round of questions, the app also attaches a short technical snapshot of that moment: the screen you were on, your screen size and pixel density, your device and browser type, which version of NavBharatAI you are running, your language, whether you were online and the connection type your browser reports, a structural note of any element that was reaching past the edge of your screen (its tag and CSS classes — never its text), and the last few error messages your browser had recorded in that tab. This snapshot is taken only when you press Send on a report, is never collected in the background, and never includes the content of your chats, your files or your apps.

2.2 Information created while you use the Platform

• Your built apps. The source files, assets and configuration of every app you build or import — stored durably so your project survives between sessions and devices.
• Build history and reports. What a build did, which steps ran, what failed and was repaired, timing, and honest success/failure verdicts. These records exist so builds can resume, so you can see what happened, and so we can fix real defects.
• Wallet and usage records. Your token balance, the ledger of purchases and deductions, and per-build/per-action usage — kept accurately because they are money.
• Preferences. Theme, text size, language, and similar settings.
• Secrets you store on purpose. API keys and credentials you save in Settings → Secrets & API Keys so your built apps can use them. These are stored encrypted, are injected only into your own apps, and are never used for any other purpose.
• Learning signals. The Platform learns from real build errors so mistakes are not repeated. Cross-user learning is anonymous by construction: it is keyed by machine "error signatures" with numbers, file paths and quoted text stripped, the stored text passes an automatic secret-and-personal-data redaction step, and no account identifier is ever attached (this design is enforced by automated tests).
• Technical logs. IP address, device/browser type, timestamps, pages visited, errors encountered. We use privacy-friendly analytics; the consent banner in the app describes and controls measurement.

2.3 Information from third parties

• Your sign-in provider (Google/Apple) shares the basic profile described above.
• If you choose to connect GitHub, we receive the repository access you grant, so we can import or save your code where you asked.
• Our payment provider tells us whether a payment succeeded, failed or was refunded.

We do not buy data about you from data brokers.

3. What we use it for

• Providing the service — answering your chats, building and previewing your apps, storing your projects, keeping your wallet accurate. (Basis: performance of the service you signed up for.)
• AI processing — sending your prompt and relevant project context to AI infrastructure so a reply or a build can be produced. (Basis: performance of the service.)
• Payments — creating orders, crediting tokens, settling incomplete orders on your next sign-in so paid money is never lost. (Basis: performance of the service; legal obligations.)
• Safety and integrity — malware-scanning Nav App Store uploads, rate-limiting abuse, blocking fraud, and an automatic check on messages and on apps before they are published. That check is software, not a person: it looks for a short list of unlawful categories (sexual content involving minors, non-consensual imagery, weapon-making instructions, drug selling). If it finds nothing — which is almost always — nothing at all is stored. If it does object, we keep the rule it matched, the time, your account id and a short extract of the message with secrets and personal identifiers stripped out, for 180 days, so a person can review it. We do not read your conversations otherwise. (Basis: legitimate interest in a safe platform, and our legal duties as an intermediary.)
• Improving the Platform — fixing defects found in honest build reports; anonymous cross-user mistake learning. (Basis: legitimate interest; anonymised data.)
• Communication — service messages (build finished, payment received, plan expiring), support replies. (Basis: performance of the service.)
• Legal compliance — tax records, responding to lawful orders. (Basis: legal obligation.)

What we do NOT do: we do not sell your personal data; we do not show third-party advertising inside NavBharatAI (we do advertise NavBharatAI itself on other platforms and measure whether those ads work — Section 3.1 sets out exactly what that shares); we do not use the private content of your chats, your uploaded documents or your built apps to train any AI model of our own or of any third party; and we do not read your projects out of curiosity — access by our team is restricted to what is needed to run the service, fix a defect you reported, or meet a legal duty.

3.2 Device check for the referral bonus (Android app only)

If you claim one of the free-credit bonuses in our Android app — for signing up, logging in with a verified email, verifying your mobile number, connecting GitHub, or applying a friend's referral code — the app checks, at that moment, that it is running on a genuine Android device. If the check cannot be completed, the signup, login and mobile bonuses are paid under the same rules as on our website, and the referral-code and GitHub bonuses wait until it can.

What that involves, precisely:

• The app reads Android's own per-app device identifier (its technical name is ANDROID_ID). It is specific to NavBharatAI on your handset, no other app can see it, and it changes if you factory-reset the device. We store it only to record that this device has claimed its bonus, so the same phone cannot claim it repeatedly.
• The app asks Google's Play Integrity service for a signed statement that this is a real device running our genuine app installed from the Play Store. That statement goes to Google and comes back to our server; it does not tell us who you are, and we do not keep it.

Why: the bonus is real money, and without this check one person could claim it endlessly from a laptop using free email addresses. We would rather protect it this way than reduce what we give you.

What this is NOT: it is not advertising, it is not shared with Meta or any advertising platform, it is not used to track you between apps or across the web, and it is never used to build a profile of you. Nothing here happens on the website, and nothing happens in the app until you actually claim a bonus.

3.3 Bot protection (Google reCAPTCHA Enterprise on the website; Play Integrity and App Attest in our apps)

When you use the NavBharatAI website, a few actions that cost us real money — building an app, sending a chat message, generating an image, asking a Professional assistant, and requesting a mobile OTP — carry a short-lived token that proves the request came from our genuine website and not from an automated script. The token is issued through Firebase App Check, using Google reCAPTCHA Enterprise.

What that involves, precisely: to issue the token, reCAPTCHA Enterprise collects information about your browser and device and how the page is being used (for example your IP address, browser and device characteristics, and interaction signals) and sends it to Google, which returns a risk judgement. Google processes this under its own Privacy Policy and Terms of Service. Our server receives only the token and whether it is valid — not the signals behind it — and we do not store the token.

Why: these actions spend money on every request, and without this check a script with a throwaway account could run them endlessly. (Basis: our legitimate interest in keeping the service secure and affordable.)

What this is NOT: it is not advertising, it is not shared with Meta or any advertising platform, and we do not use it to build a profile of you. In our apps it works differently: the Android app asks Google's Play Integrity service, and the iOS app asks Apple's App Attest service, to confirm the app is genuine and unmodified. That check is about the app and the device, not about you — it sends no chats, files or contact details, and our server again receives only the resulting token and whether it is valid.

3.4 Crash reporting in our apps (Google Firebase Crashlytics)

When the NavBharatAI Android or iOS app crashes or hits an error, the app sends a crash report to Google Firebase Crashlytics so we can find and fix the cause. A report contains: the error and where in our code it happened; the app version and build number; the device model and operating-system version, with technical state such as free memory and disk space; the time; which screen of the app you were on; whether you were signed in; and, if you were, a one-way scrambled form of your account ID, so that repeated crashes for one person can be grouped without revealing who they are. Firebase also assigns the installation a random identifier.

What a crash report never contains: never your chats, prompts, files, email address, phone number, payment details or passwords. Every report is passed through a filter on your device that removes tokens, keys, card numbers, UPI addresses, email addresses, phone numbers and web-address parameters before it is sent. Google processes crash reports under its own Privacy Policy and keeps them for 90 days.

Why: an app that crashes for some people and not others cannot be fixed without knowing where it crashed. (Basis: our legitimate interest in keeping the apps working.) On the website, the same filtered error details are sent only to our own servers, as part of the technical logs in Section 2.2.

3.1 Advertising measurement (Meta / Facebook and Instagram)

We advertise NavBharatAI on Facebook and Instagram so people can find it. To know which of those ads actually bring people — rather than guessing and wasting money — we share a small, fixed set of events with Meta.

On the website, if and only if you accept the consent banner, a Meta pixel loads and reports:

• that a page was viewed;
• that an account was created;
• that a purchase completed, with the real amount in rupees;
• that an app was built.

In our Android app, again only if you accept the consent banner, builds that include Meta's measurement SDK report that the app was installed or opened, together with your device's advertising ID — the resettable identifier Android provides for exactly this purpose. Until you accept, that SDK is not switched on at all.

What is NEVER shared with Meta or any advertising platform:

• the content of your chats or prompts;
• files or documents you upload;
• anything from the Doctor AI / clinical surface;
• your built apps, their code, or their data;
• your name, email address or phone number.

This is enforced in our code, not just promised: the pixel can only send the four events listed above, because that list is written into the software as a fixed allowlist.

Your control. On the web, choose Decline on the consent banner and no advertising measurement runs at all — nothing loads, nothing is sent. The Android app honours the same choice: the measurement SDK ships switched off and is turned on only after you accept, and turned back off if you later withdraw. Separately, your phone's Settings → Privacy → Ads lets you reset or delete your advertising ID at any time.

(Basis: your consent, which you can withdraw at any time.)

4. AI processing — an honest explanation

NavBharatAI's intelligence is delivered by NavBharatAI's engine, which runs on enterprise AI infrastructure. What you should know:

• Your prompts, relevant project files and conversation context are processed by AI infrastructure providers acting as our processors, only to generate the reply or build you asked for.
• We contractually require that this data is not used to train those providers' models.
• AI output can be wrong. Verify important results — and see the specific medical disclaimer in Section 5.
• Provider-level processing details are available to business customers under our Data Processing Agreement, including a sub-processor description by category.

5. Sensitive data: the Doctor AI / clinical surface (web)

The Senior Doctor Assistant is a clinical decision-support tool intended for qualified doctors. All diagnoses and treatment decisions remain the sole responsibility of the treating physician. It is not available in our Android app.

If you use it, you may enter health-related information or upload clinical documents. For this data specifically:

• It is used only to produce the assistance you asked for in that case, and to keep your own case history available to you.
• Cases are isolated per case-session, so a new patient's case never inherits a previous patient's context.
• It is never used for advertising, profiling, or model training.
• If you are a doctor entering information about a patient, you are responsible for having the patient's consent and for entering only what is needed. Prefer de-identified details wherever possible.
• You can delete a case, and account deletion removes your case history (Section 9).

6. Where your data lives, and for how long

• Hosting. The Platform runs on enterprise-grade cloud infrastructure. Our primary serving and storage region is Asia (Singapore), with some services in other regions of the same providers. Because Singapore is outside India, your data is transferred across borders; we protect it in transit with TLS encryption and at rest with provider-grade encryption, and we transfer it only where permitted by applicable law.
• Retention. We keep your account data, projects, chats and wallet records for as long as your account is active, because the product's promise is that your work and history are there when you return. Specific shorter windows: build sandboxes are recycled automatically after inactivity (minutes, not days); the text of the latest document you attach in a NavBharatAI Pro chat (with personal identifiers such as phone and Aadhaar numbers masked) is kept for that one chat for up to 30 days, so a later message in the same chat can use it — it is never used in any other chat, and it is deleted with the chat or when you unsend the message; technical logs are retained for up to 90 days; the reports of your past builds (what you asked for, what the build engine did, and what failed) are kept for 180 days so a defect can be traced, and are deleted sooner with the project or your account; payment and tax records are retained for the period Indian law requires (typically 8 years) even after account deletion.
• Safety-check records. The flagged-message records described in Section 3 are kept for 180 days and then deleted. They survive account deletion for that period — a record of abuse that can be erased by closing the account is not a record.
• App Mart notifications and comment reports. App Mart notifications ("Asha liked your app") are deleted after 90 days. If someone reports a comment, we keep the report — the reason, a copy of the reported comment's text, and who reported it — for 180 days so a person can review it, and like the other safety records it survives the deletion of either account for that period.
• Removal records. If we remove something you published — an app, a listing, a site — we keep a short record of what was removed, why, when, and who published it, for 180 days, because the Information Technology (Intermediary Guidelines) Rules, 2021 require it for investigation. That record holds a description and a fingerprint of the content, never a copy of the content itself, and it is used only for that purpose. See our Grievance Redressal page.
• After deletion. When you delete your account (Section 9), personal data is deleted or irreversibly anonymised within 30 days, except the payment/tax records above, the 180-day safety-check and removal records just described, and data we must keep for a live legal matter.
• Apps you published to the Nav App Store. These are anonymised rather than deleted. Your display name, your public creator code and the link to your account are removed, and the listing is taken out of the public store — but the app itself is kept, because people who bought it would otherwise lose what they paid for, and a purchase is not refundable. Once your account is gone a paid app of yours can no longer be bought, since there would be nobody to pay. Submissions that were never published are deleted outright. If you want a published app taken down completely, ask us and we will remove it.

7. Who we share data with

We share personal data only with the parties below, and only for the purposes described:

• Payment processing: Cashfree Payments — processes your payments; we exchange order identifiers and outcomes with them. Their handling of your card/UPI data is governed by their own policy and RBI regulations.
• Sign-in: Google / Apple — when you choose them as your login method.
• Cloud infrastructure providers — host our servers, databases and file storage under strict contracts.
• AI infrastructure providers — process prompts/context to generate output, as described in Section 4, with no training rights.
• GitHub — only if you connect it, and only with the access you granted.
• Bot protection: Google reCAPTCHA Enterprise, Google Play Integrity and Apple App Attest (via Firebase App Check) — for the actions listed in Section 3.3.
• Crash reporting: Google Firebase Crashlytics — the filtered crash reports described in Section 3.4, from our Android and iOS apps only.
• Malware scanning — files uploaded to the Nav App Store are submitted to an anti-malware scanning service before any listing can be approved.
• Meta (Facebook / Instagram) — the advertising-measurement events listed in Section 3.1, and only with your consent. Never your chats, files, clinical data or built apps.
• Authorities — if required by a valid legal order. We check every demand and share the minimum required.
• A future business transfer — if NavBharatAI is ever acquired or merged, data transfers with the business; this policy (or one no less protective) continues to apply, and we will notify you.

We do not sell your personal data and we never share it with data brokers. The only advertising-related sharing we do is the conversion measurement described in Section 3.1 — a fixed list of events, only after you consent, and never the content of anything you write, upload or build.

8. Security

• Encryption in transit (TLS) everywhere, and at rest on our storage systems.
• Secrets you store are encrypted with a dedicated key separate from ordinary application data.
• Automatic redaction of secrets and personal identifiers (API-key shapes, emails, phone numbers, Indian identifiers such as PAN/Aadhaar patterns) from logs, tool output shown on screen, and the anonymous learning system.
• Strict access control inside the team: production access is limited, logged, and need-based.
• Every app uploaded to the Nav App Store is signature-checked and malware-scanned, and nothing publishes without passing review — a failed or unavailable scan blocks publication rather than being skipped.
• Independent security research is welcome — see Security at NavBharatAI for the full detail: encryption, access control, incident response, and our vulnerability disclosure process.

No system on the internet is perfectly secure. If a breach affecting your personal data occurs, we will notify the affected users and the authorities as the DPDP Act requires, without undue delay.

9. Your rights and controls

Under the DPDP Act (and simply because it is right), you can:

• Access — ask what personal data we hold about you.
• Correct — fix inaccurate account data (much of it is editable in Settings directly).
• Delete — delete individual projects/chats in-app, or request full account deletion by email. Deletion follows the timelines in Section 6.
• Withdraw consent — for anything based on consent (e.g. optional analytics via the consent banner), withdraw it at any time without losing the core service.
• Grievance — raise a complaint at info@navbharatai.com, or through our Grievance Redressal page, which names the officer responsible and the timelines we must answer within under the IT Rules, 2021. If you are unsatisfied with our response, you may escalate to the Data Protection Board of India.
• Nominate — under the DPDP Act you may nominate a person to exercise these rights for you in case of death or incapacity; email us to record a nomination.

We will verify identity before acting on a request, so nobody else can use these rights against you.

10. Children

The Platform is not intended for children under 18. We do not knowingly collect personal data from children. If you are a parent or guardian and believe a child has created an account, contact us and we will delete it. Where verifiable parental consent becomes practical to support under DPDP rules, we may allow supervised use; until then, under-18 use is not permitted.

11. Cookies and similar technologies

We use a small set of cookies and local-storage keys that are necessary for the Platform to work (your session, your theme, your text-size choice, your consent decision). Optional analytics and the Meta advertising-measurement pixel (Section 3.1) run only if you accept the consent banner, which is where you control them. Decline, and no third-party advertising or measurement cookie is set at all — the pixel is not even downloaded.

Using NavBharatAI without signing in. A visitor who has not signed in may send a limited number of free AI messages each day. So that we can count them, your browser or the app keeps a random guest identifier in local storage and sends it with AI requests only while you are signed out. It is not linked to your name, phone number or any account, and it is never sent once you sign in. On our side it is stored only as a one-way code beside a daily message count, and that count is deleted after a few days. Clearing your site data removes the identifier.

11.1 Counting visits to NavBharatAI itself — no cookie, no script, no profile

So that we can tell whether anyone is actually using NavBharatAI, our own server counts the pages it serves and the times the mobile app is opened. This happens on our side, from the request your browser or the app already sends us. It sets no cookie, stores nothing on your device, downloads no script, and does not follow you anywhere.

What one count records: the page path (never the query string), whether the request came from the website or from the mobile app, and a short code derived from your IP address and browser using a secret that rotates every day — so you are counted once a day, the code cannot be turned back into you, and today's code cannot be matched to yesterday's. That daily rotation is why we can tell you how many people visited today and can never produce an all-time list of people: the ability to recognise you tomorrow is deliberately not kept.

Requests that identify themselves as bots are not counted, and browsers that send Do Not Track or Global Privacy Control are not counted at all. These counts are seen only by NavBharatAI's own administrators, never shared, and never joined to your account, your chats, your files or your apps. Counting how many people came is not the same as knowing who they were, and we do not do the second.

12. Apps YOU build, and their users

When you build an app with NavBharatAI and share or deploy it, you are responsible to the people who use it. Their data lives in the database and services you connect (your own Supabase/Firebase/other account — NavBharatAI deliberately provisions user apps on the builder's own accounts, not ours). If your app collects personal data, you need your own privacy policy for it — you can generate a starting draft with NavBharatAI, and the same review advice applies: have it checked before you rely on it.

Visitor counts for apps published on NavBharatAI hosting. So that you can see whether anyone is visiting the app you published, pages served from NavBharatAI's own hosting carry a small script that reports each visit to us. It sends only: the page path (never the query string), the referring site's hostname (never the full address), how many bytes the page and its files transferred (a number the browser itself measures, so we can show you your app's traffic and account for our own hosting costs — it describes the page, not the person), and a short code derived from the visitor's IP address and browser using a secret that rotates daily — so the same visitor is counted once per day, the code cannot be turned back into a person, and it cannot be joined across days. It sets no cookie and stores nothing on the visitor's device. Browsers that send Do Not Track or Global Privacy Control are not counted at all. These counts are shown only to you, the app's builder, and are kept for 30 days. Apps you deploy to your own hosting provider do not carry this script.

13. Changes to this policy

When we change this policy in a meaningful way, we will update the date at the top and announce the change in-app before it takes effect. Continued use after the effective date means the updated policy applies. The current version always lives at this page.

This policy is provided in English as its authoritative version. Summaries in other languages may be offered for convenience; if they conflict, the English version controls.