Privacy Policy
Version 0.1.0-draftUpdated 24 October 2026
Draft — under legal review. Not legal advice and not yet in force.
Highlighted notes like [VERIFY: …] and blanks still to be filled in show this is a working draft.
On this page
- In short (plain language)
- 1. Who we are
- 2. Laws that apply
- 3. Scope
- 4. Personal data we collect
- 5. Why we use it (purposes) and on what basis
- 6. Consent — how we ask and how you withdraw
- 7. Third-party AI processing (Google Gemini)
- 8. Who we share personal data with
- 9. Where your data is stored, and transfers outside India
- 10. How long we keep data
- 11. Deleting your account
- 12. Security
- 13. Your rights and how to use them
- 14. Your duties
- 15. Children
- 16. Personal data breaches
- 17. Cookies and similar storage
- 18. Automated decisions
- 19. Changes to this policy
- 20. Contact
Version 0.1 (draft) · Effective: {{EFFECTIVE_DATE}} · Last updated: 24 October 2026 · Notice version: privacy-2026-10-v0.1
In short (plain language)
- Who: Trinetra AI Labs Private Limited runs DamruVani and decides how your personal data is used (we are the "Data Fiduciary").
- What we collect: your name and email, the text and audio you give us, the audio we generate for you, voice recordings if you create a voice clone, and technical data needed to keep the service secure. If you choose to: waitlist details, invite / referral links, a phone number for verification, survey answers and feedback, and your number if you join our WhatsApp group. Billing details only once paid plans exist.
- Why: to run the service, keep it safe, prevent abuse and meet legal duties. We do not sell your data and we do not show ads.
- Your voice is protected: we clone a voice only with your separate, explicit consent and a live check. We do not use your data — or any audio generated with Model 1 — to train AI models.
- Third-party AI (Google Gemini) is switched off during the beta. If we switch on features such as auto-tagging emotions or fixing Hinglish spelling, some text will go to Gemini — we will update this policy before that happens (§7).
- Where: our servers and GPU are planned to be in India. Encrypted backups are stored with Cloudflare R2 outside India, and a few providers (email, error monitoring) may process data outside India.
- How long: uploads about 7 days, generated audio for your plan's history period, account data until you delete your account. Billing records must be kept for about 8 years under tax and company law.
- Your rights: see, correct and delete your data, withdraw consent, nominate someone, and complain — from Settings → Privacy or by writing to our Grievance Officer at grievance@{{DOMAIN}}. We reply within 30 days.
- Only 18+. Cookies: strictly necessary only — no tracking or advertising cookies.
1. Who we are
| Data Fiduciary | Trinetra AI Labs Private Limited (CIN U46101GJ2025PTC166089) [VERIFY: name spelling + address from MCA] |
| Registered office | {{REGISTERED_OFFICE_ADDRESS}} |
| Service | DamruVani — website {{DOMAIN}}, Studio, API (api.{{DOMAIN}}), mobile apps |
| Contact for privacy questions and rights requests (DPDP §8(9)) and Grievance Officer | {{GRIEVANCE_OFFICER_NAME}}, {{GRIEVANCE_OFFICER_DESIGNATION}} — grievance@{{DOMAIN}} — {{REGISTERED_OFFICE_ADDRESS}} — see /legal/grievance |
2. Laws that apply
- Digital Personal Data Protection Act, 2023 ("DPDP Act") and the Digital Personal Data Protection Rules, 2025 ("DPDP Rules"), notified in November 2025. Most duties of data fiduciaries under the Rules apply from 13 May 2027. [VERIFY: commencement dates against the Gazette notification (G.S.R. 846(E)); a MeitY proposal to bring them forward was reported in early 2026 but, per trackers, had not been notified as of October 2026] We follow the DPDP standard now.
- Until the DPDP Act fully replaces it, section 43A of the Information Technology Act, 2000 and the Information Technology (Reasonable Security Practices and Procedures and Sensitive Personal Data or Information) Rules, 2011 ("SPDI Rules") also apply. Under the SPDI Rules, biometric information — which includes voice patterns — is sensitive personal data. [VERIFY: SPDI Rule 2(1)(b) wording] [LAWYER: confirm voice-clone consent meets SPDI Rule 5]
- The IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, as amended in 2026 ("IT Rules"), and the CERT-In Directions of 28 April 2022 also apply to parts of our processing. [LAWYER: whether DamruVani is an "intermediary"]
3. Scope
This policy covers personal data we process when you use the DamruVani website, Studio, API and mobile apps. The Service is for people in India who are 18 or older. If you use the API to build your own application, you are responsible for your own users' personal data and must give them your own privacy notice.
4. Personal data we collect
| Category | Examples | Where it comes from | Needed? |
|---|---|---|---|
| Account | Name, email address, Google account identifier (if you sign in with Google), the Terms and notice versions you accepted and when, your 18+ declaration | You; Google (only if you choose Google sign-in) | Yes |
| Preferences | UI language, main language, use-case and team size (onboarding — you can skip), favourite voices | You | Optional |
| Security and technical | IP address, browser / device type, app version, session and sign-in records, bot-protection signals (Cloudflare Turnstile), on Android a hashed install identifier and a Play Integrity verdict | Your device; Cloudflare; Google Play | Yes — for security and abuse prevention |
| Content (Input) | Text and scripts you type or upload, audio / video you upload for transcription or dubbing, pronunciation-dictionary entries | You | Yes, for the feature you use |
| Output | Audio, transcripts and subtitles generated for you; job details | Generated by the Service | Yes |
| Voice data (only if you clone a voice) | Your voice samples, your live challenge-sentence recording and its transcript and score, the derived voice model ("voice prompt") and speaker-similarity measurements, consent record (time, consent text version, IP address, browser) | You; generated by the Service | Only for Voice Lab |
| Usage | Credits used, job counts and types, per-voice usage, limits reached | Generated by the Service | Yes |
| Communications | Support tickets, abuse reports and grievances, email delivery status | You; our email provider | When you contact us |
| Billing (only after paid launch) | Billing name, address and state, GSTIN (if any), invoices, payment and refund records (payment ID, method type, amount) | You; Razorpay | When you pay |
| Mobile push (Android) | Push token | Your device / Firebase | If you allow notifications |
| Phone number (optional) | Mobile number you verify for a one-time bonus and to limit duplicate free accounts; verification-code delivery status | You; our SMS provider | Optional [OWNER: only if phone verification ships in the Beta] |
| Waitlist and invitations | Email, name, main language, intended use, social / website link (optional), how you heard of us (campaign tag / UTM parameters), our internal priority score, invite-code status | You; the link you arrived from | If you join the waitlist |
| Referrals | Which invite code you used or shared (who invited whom), whether an invited person activated their account, referral credits | Generated by the Service | If you use or share an invite code |
| Community groups | If you join our WhatsApp Channel or beta WhatsApp group: your phone number and WhatsApp profile name. In a group, other members can see your number; WhatsApp (Meta) processes it under its own terms | You; WhatsApp | Optional — you choose to join |
| Feedback and research | 👍 / 👎 ratings and reasons, survey answers (for example satisfaction, product fit, interest in paid plans), notes from feedback calls you agree to | You | Optional |
We do not collect government ID numbers, contacts, location, the content of your SMS messages or your full card / UPI details. Please do not put sensitive personal data (for example health, financial or ID details of yourself or others) into text you convert to speech.
5. Why we use it (purposes) and on what basis
| Purpose | Data used | Basis | Default |
|---|---|---|---|
| Create and run your account; sign you in | Account, security | Consent at sign-up (DPDP §6) [LAWYER: or legitimate use, §7(a)] | Required |
| Provide the features you use: generate speech, transcribe, dub, store your history for your plan's period | Content, Output, usage | Consent at sign-up | Required for that feature |
| Voice cloning | Voice data | Separate, explicit consent in Voice Lab (Voice Clone Consent) | Off until you create a clone |
| Keep the Service secure; prevent fraud, abuse and multiple free accounts; moderate content | Security, content, usage | Consent at sign-up; compliance with law [LAWYER: map to DPDP §7 legitimate use vs consent] | Required |
| Provenance — label Output and keep a record linking each file's fingerprint to the job that made it | Output fingerprint, job and account IDs | Compliance with law (IT Rules); consent at sign-up | Always on |
| Credits, limits, invoices and tax records | Usage, billing | Contract / consent; legal obligation (GST and company law) — DPDP §7 [LAWYER: exact clause] | Required when applicable |
| Support and grievances | Communications | Consent; legal obligation (grievance redressal) | When you contact us |
| Waitlist, beta invitations and selection, referral rewards | Waitlist, referrals, account | Consent (when you join the waitlist or use an invite code) | When you join |
| Phone verification bonus and limiting duplicate free accounts | Phone number | Consent | Optional |
| Product research: surveys, feedback calls, ratings | Feedback and research | Consent | Optional |
| Community announcements and beta group | Community groups | Consent (you choose to join; WhatsApp acts under its own terms) | Optional |
| Training or improving speech models | Not offered during the Beta. We do not use Content, Output or voice data for training. Audio generated with Model 1 is never used for training (its licences forbid it). If we ever offer a programme to contribute your own recordings, we will ask separately first and update this policy | Separate opt-in consent (only if offered) | Not offered |
| Product news, offers and marketing emails | Email, name | Separate opt-in consent | OFF |
| Legal requests, disputes, enforcement of our Terms | Any relevant data | Compliance with law | As needed |
We do not sell personal data, use it for advertising, or build advertising profiles.
6. Consent — how we ask and how you withdraw
6.1 We ask for consent separately for each purpose that needs it, in clear language, with an itemised notice. We record which notice version you saw and when.
6.2 Withdrawing is as easy as giving consent: Settings → Privacy has a switch for each optional purpose (for example marketing emails; a phone number can be removed there too). Deleting a Voice Clone withdraws voice-clone consent. Every marketing email has an unsubscribe link. You can leave a WhatsApp group or Channel at any time.
6.3 Withdrawal stops future processing for that purpose. It does not affect processing already done. If you withdraw consent needed for a core feature, we may not be able to provide that feature.
6.4 Model training: we do not use your data for training during the Beta. If we ever offer an opt-in programme, we will explain before you join that information used to train a model cannot be removed from a model that has already been trained, and that withdrawing keeps your data out of all future training.
6.5 When India's registered Consent Managers become available, we may also let you manage consent through one. [VERIFY: Consent Manager provisions commence 13 November 2026]
7. Third-party AI processing (Google Gemini)
Status: switched off during the Beta — no text is sent to Gemini. [OWNER: keep in sync with TEXT_AI_PROVIDER at publish time] The table below describes what will happen if we switch a feature on; we will update this policy, its notice version and the processors table before that, and tell you in the Service.
When switched on, some text features send text to Google's Gemini API (paid tier):
| Feature | When text is sent | Can you avoid it? |
|---|---|---|
| Auto-tag emotions | Only when you switch auto-tag on | Yes — keep auto-tag off |
| Hinglish / mixed-script fix | Automatically, when we detect Roman-script Hindi (or similar mixed text) in your input | {{GEMINI_OPT_OUT_STATUS}} [OWNER: add a "Don't send my text to third-party AI" switch — quality may drop] |
| "Convert to dialect" | Only when you use it and approve the result | Yes |
| Translation (dubbing, transcript translation) | Only when you use it | Yes |
| Safety moderation | [VERIFY: whether moderation will use Gemini or another provider] | No — needed for safety |
- We send only the text needed for the task, plus the list of allowed emotion tags. We do not add your name, email or account ID. The text itself may still contain personal data if you typed it — please avoid that.
- Under Google's paid-tier terms, Google does not use these prompts or responses to improve its products; it may keep them for a limited period to detect abuse, and they may be processed in any country where Google has facilities. [VERIFY: re-check Gemini API Additional Terms at launch]
- Gemini processes data for us as our Data Processor. Results are checked by our own validators; your original words are never changed without your approval.
8. Who we share personal data with
8.1 Data Processors — providers who process personal data only on our instructions, under contracts that require reasonable security safeguards (DPDP Rules, Rule 6):
| Provider | What they do for us | Personal data involved | Location | Status |
|---|---|---|---|---|
| {{VPS_PROVIDER}} (cloud server) | Hosts the website, API, database, job queue and file storage | All categories | India region planned [VERIFY at purchase] | Not yet chosen |
| Our own GPU server | Generates speech and transcripts | Text segments, uploaded audio, voice prompts, temporary Output | Operated by us in Gujarat, India | In use (beta) |
| Cloudflare, Inc. | DNS, content delivery, DDoS / firewall protection, bot protection (Turnstile); planned object storage (R2) for files and public voice previews | IP address, request data, Turnstile device signals; stored files if R2 storage is adopted | Global network; R2 location [VERIFY: R2 has no India-only location — check location hints at setup] | Planned |
| Cloudflare, Inc. — R2 (backups) | Stores our encrypted database backups and weekly voice-reference backups. The decryption key is held only by us — Cloudflare cannot read them | All categories in the database (encrypted); voice references (encrypted) | Outside India (Asia-Pacific location hint) [VERIFY at setup] | In use from launch |
| {{SMS_PROVIDER}} (SMS verification) | Sends phone verification codes | Phone number, code delivery status | [VERIFY: provider + location; Indian SMS rules (TRAI DLT registration)] | Only if phone verification ships |
| Resend (primary) / Amazon Web Services SES, Mumbai (fallback) | Sends sign-in codes and service emails | Email address, email content, delivery events | Resend: United States [VERIFY: Resend docs say account data is stored in the US whatever the sending region]; SES: Mumbai, India | Planned |
| Google LLC — Gemini API | Text features listed in §7 | Text you submit | May be outside India | Planned (Phase 1) |
| Functional Software, Inc. (Sentry) | Error and crash monitoring for web, API and Android app | Technical error data, device / browser details, pseudonymous account ID; we filter out email and text content | United States or EU (chosen at setup) [VERIFY: region is permanent once chosen — decide before creating the account] | Planned |
| Tailscale Inc. | Encrypted private network between our servers | Network metadata only (device names, network addresses) — job content travels end-to-end encrypted | [VERIFY: Tailscale coordination-server location and data list] | Planned |
| Google — Firebase Cloud Messaging | Android push notifications | Push token; notifications carry no personal content | Global | Android app |
| {{OPS_ALERT_CHANNEL}} (e.g. Telegram) | Internal operational alerts to our team | Masked identifiers only, no customer content [VERIFY: alert payloads are masked] | [VERIFY: provider's server location] | Planned |
8.2 Independent parties that act as data fiduciaries in their own right, under their own privacy policies: Google (if you sign in with Google), Razorpay (payments, once paid plans exist; Razorpay test mode during the Beta), banks and UPI apps, Google Play (app distribution and Play Integrity), and WhatsApp / Meta Platforms (if you join our WhatsApp Channel or group). [LAWYER: Meta's role for a business-run group]
8.3 Legal disclosures. We disclose personal data when required by law — for example to courts, government agencies and law enforcement acting under valid legal authority, or to CERT-In. Under the IT Rules, we may disclose the identity of a user who violated the rules to a complainant who is a victim, in accordance with applicable law. We record every such request.
8.4 Business transfers. If our business is merged or sold, personal data may transfer to the new owner, who must respect this policy. We will tell you first. [LAWYER: DPDP has no explicit merger clause — confirm notice / consent needs]
8.5 We do not share personal data with advertisers or data brokers.
9. Where your data is stored, and transfers outside India
9.1 We plan to keep our main servers and database in India, and our GPU server is in India. [VERIFY: when the cloud server is bought]
9.2 Some processors may process personal data outside India: our encrypted backups in Cloudflare R2, Google (Firebase; Gemini only if switched on), the email provider, the error-monitoring provider, and Cloudflare's global network. Under section 16 of the DPDP Act, transfers are allowed except to countries the Central Government restricts by notification; we will stop any transfer that becomes restricted. [VERIFY: no country notified as of 24 October 2026] [LAWYER: SPDI Rule 7 requires the same level of protection, and transfer only where necessary or with consent]
10. How long we keep data
| Data | Kept for | Then |
|---|---|---|
| Uploaded files (input for transcription, dubbing, cloning uploads) | 7 days after the job; 24 hours if the upload is not completed | Deleted |
| Generated Output and history | Your plan's history period (Free: 7 days; paid plans as shown on the Pricing page) | Deleted |
| Temporary files on our GPU server | Deleted right after each job (at most 24 hours) | Deleted |
| Job queue records | 24 hours (completed) / 7 days (failed) | Deleted |
| Voice Clone data (samples, voice prompt, challenge recording) | Until you delete the clone or your account | Deleted, including from GPU caches |
| Voice-clone consent record (text version, time, IP — no audio) | {{CONSENT_EVIDENCE_RETENTION}} after the clone is deleted [LAWYER: proof of consent (DPDP §6(10)) vs. erasure; proposal: 3 years] | Deleted |
| Account profile | Until you delete your account (plus a 7-day cancellation window) | Deleted or anonymised |
| Provenance record (Output fingerprint → job → account → voice) | 1 year | Deleted; after account deletion kept without personal identifiers |
| Sign-in and security events | 180 days (IP addresses stored hashed) | Deleted |
| Email delivery logs | 180 days | Personal data masked |
| Administrative audit logs | 3 years (billing-related entries 8 years); contain no personal data beyond IDs | Deleted |
| Billing records — invoices, payments, credit notes, refunds, billing profile at time of invoice | About 8 years: 72 months from the due date of the annual GST return for the year (CGST Act §36), and at least 8 financial years for books of account (Companies Act §128(5)) [CA: confirm] | Deleted |
| Content removed after a valid complaint (takedown) | 30 days, unless law enforcement or a court asks us to preserve it | Deleted |
| Support tickets and grievance records | {{SUPPORT_RETENTION}} [OWNER: proposal 2 years after closure] | Deleted |
| Abuse reports and enforcement records | {{ABUSE_RETENTION}} [OWNER/LAWYER: proposal 3 years] | Deleted |
| Backups (encrypted, Cloudflare R2 outside India) | Rolling 30 days (database daily; voice references weekly) | Deleted |
| Waitlist entries | Until you join, or {{WAITLIST_RETENTION}} after our last contact [OWNER: proposal 12 months] | Deleted |
| Referral records | While both accounts exist; then anonymised | Anonymised |
| Phone number | Until you remove it or delete your account; delivery logs 180 days | Deleted |
| Feedback, survey answers and call notes | {{RESEARCH_RETENTION}} [OWNER: proposal 2 years]; quotes are published only with your separate permission | Deleted or aggregated |
| Web server access logs (full IP address, browser) | 180 days (security; CERT-In Directions) | Deleted |
| Logs and personal data needed for security investigations | At least 1 year where the DPDP Rules require it (Rule 6(1)(e), from 13 May 2027 [VERIFY: Gazette text + commencement]); system logs 180 days in India (CERT-In Directions [VERIFY: current Directions text]) | Deleted |
Text typed into the public "try it" box on our website is not stored. [OWNER: RULES R-D-09 is still "proposed" — confirm it is implemented before publishing]
11. Deleting your account
11.1 Delete your account from Settings → Privacy, the mobile app, or /account/delete. There is a 7-day grace period in which you can cancel.
11.2 After that we delete your profile, Voice Clones (including copies cached on our GPU server), consent audio, uploads, Outputs, projects, API keys and sessions, and cancel any subscription.
11.3 What we must keep: invoices, payment, refund and credit-note records and the billing profile used on them, for the period tax and company law requires (§10). Your credit history is anonymised. Copies in backups disappear within 30 days.
11.4 We email you when deletion is complete.
12. Security
We protect personal data with reasonable security safeguards, including: encryption in transit (HTTPS / TLS) and for backups; databases, queues and storage not exposed to the public internet; private, encrypted networking to our GPU server; access controls with two-factor authentication for staff and a tamper-evident audit log; hashing of API keys and sign-in codes, and of IP addresses in our sign-in and security records (our web server access logs keep full IP addresses and browser details for 180 days for security, as Indian cyber-security directions require); masking of email, phone and text in application logs; short-lived signed links for file downloads; isolated, sandboxed processing of uploaded media; and automated security scans and tests before releases. No system is perfectly secure; we will tell you about a breach as described in §16. [LAWYER: SPDI Rule 8 refers to IS/ISO/IEC 27001 or an approved code — we are not certified]
13. Your rights and how to use them
| Right | What it means | How | Response time |
|---|---|---|---|
| Access (DPDP §11) | A summary of your personal data, how we process it, and who we shared it with | Settings → Privacy → Export my data (download link sent to your email), or write to grievance@{{DOMAIN}} | Within 30 days |
| Correction, completion and update (§12) | Fix inaccurate or incomplete data | Settings → Profile / Billing; or write to us. Issued tax invoices can only be corrected through credit notes or re-issue, as tax law allows | Within 30 days |
| Erasure (§12) | Delete data no longer needed or for which you withdraw consent | Delete individual items, a Voice Clone, or your whole account (§11); or write to us | Grace period + deletion as in §11; legal records kept as in §10 |
| Withdraw consent (§6(4)) | Stop optional processing | Settings → Privacy switches; delete a Voice Clone; unsubscribe link | Immediate for future processing |
| Grievance redressal (§13) | Complain to us about how we handle your data | /legal/grievance | Within 30 days |
| Nominate (§14) | Name someone to exercise your rights if you die or become unable to | Write to grievance@{{DOMAIN}} [OWNER: build a nomination form] | Within 30 days |
| Complain to the Data Protection Board | If you are not satisfied after using our grievance process | Data Protection Board of India — {{DPB_COMPLAINT_URL}} [VERIFY: official complaint channel] | — |
- How we identify you: to protect your account, we act on requests made while you are signed in (we may ask you to sign in again), or from your registered email address after confirming a one-time code. The identifier we need is your registered email address.
- We aim to respond within 30 days — faster than the 90-day maximum allowed by the DPDP Rules (Rule 14(3)) [VERIFY: Gazette text; applies from 13 May 2027], and within the one-month period of SPDI Rule 5(9).
- Rights may be limited where the law requires us to keep or disclose data.
14. Your duties
The DPDP Act (§15) asks you to: follow applicable law when exercising your rights; not impersonate another person when giving us information; not suppress material information when giving us personal data for an official document; not file false or frivolous complaints; and provide only verifiably authentic information when you ask for correction or erasure. [VERIFY: penalty for breach of these duties under the DPDP Act Schedule]
15. Children
The Service is only for people aged 18 or older. We do not knowingly collect personal data from children, and we do not allow cloning of a child's voice. If we learn that a user is under 18, we will suspend the account and delete their personal data, except what the law requires us to keep. If you believe a child is using DamruVani, tell us at grievance@{{DOMAIN}}.
16. Personal data breaches
If a personal data breach affects you, we will tell you without delay — by email and in the Service — in plain language: what happened (nature, extent, timing), the likely consequences for you, what we are doing to reduce the risk, steps you can take to protect yourself, and whom to contact. We will also inform the Data Protection Board as the DPDP Rules require (an initial report without delay and a detailed report within 72 hours, from 13 May 2027 [VERIFY: Rule 7 text + commencement]), and report cyber-security incidents to CERT-In within the time its Directions require (6 hours from noticing the incident [VERIFY: CERT-In Directions of 28 April 2022, still current?]).
17. Cookies and similar storage
We use only strictly necessary cookies and browser storage (sign-in, security, remembering small preferences). No advertising or tracking cookies, and no third-party analytics. Details: Cookie Policy.
18. Automated decisions
- Our moderation systems may automatically refuse a generation that appears to break the AUP.
- Voice-clone verification compares your challenge recording with the expected sentence and with your sample automatically; our team reviews some attempts manually.
- These checks do not produce legal effects such as credit or employment decisions. You can ask for a human review of any automated block through /legal/grievance.
19. Changes to this policy
When we change this policy we update the version and date at the top. For material changes we notify you by email or in the app before they take effect, and where the law requires it we ask for your consent again. Earlier versions are archived at /legal/privacy/versions.
20. Contact
Grievance Officer: {{GRIEVANCE_OFFICER_NAME}}, {{GRIEVANCE_OFFICER_DESIGNATION}}, Trinetra AI Labs Private Limited, {{REGISTERED_OFFICE_ADDRESS}} — grievance@{{DOMAIN}}. Timelines and escalation: /legal/grievance.