Terms of Service

Version 0.1 (draft) · Effective: {{EFFECTIVE_DATE}} · Last updated: 24 October 2026

In short (plain language)

• DamruVani is an AI voice service for people in India — text-to-speech with emotions, speech-to-text, dialogue, dubbing and an API. It is run by Trinetra AI Labs Private Limited.
• You must be 18 or older to use it.
• We are in public beta. The beta is free and for non-commercial use only. The speech engine used in the beta ("Model 1") is built on third-party model weights that are licensed for non-commercial use only. Commercial use will come only with our own model ("Model 2") and paid plans — and audio you make during the beta stays non-commercial even after that.
• Everything DamruVani generates is AI-generated. We label it, and you must not remove those labels or pass the audio off as a real person's recording in a way that could mislead anyone.
• Only clone your own voice. Never use DamruVani to impersonate anyone, cheat people, make political deepfakes, or create sexual content or content involving children. The Acceptable Use Policy has the full list.
• Credits are a usage allowance, not money.
• The Service runs on limited hardware and is best-effort — there is no uptime guarantee.
• Indian law applies. Questions or complaints: our Grievance Officer at grievance@{{DOMAIN}} (/legal/grievance).

1. About these Terms

1.1 These Terms of Service ("Terms") are a contract between you and Trinetra AI Labs Private Limited, a company incorporated in India (CIN U46101GJ2025PTC166089) with its registered office at {{REGISTERED_OFFICE_ADDRESS}} ("DamruVani", "we", "us", "our"). [VERIFY: exact name spelling and registered address from MCA master data]

1.2 These Terms apply to the DamruVani website at {{DOMAIN}}, the DamruVani Studio web app, the public API at api.{{DOMAIN}}, our mobile apps, and every related feature (together, the "Service").

1.3 The following policies form part of these Terms: the Privacy Policy, the Acceptable Use Policy ("AUP"), the Voice Clone Consent, the Cookie Policy, the Refund and Cancellation Policy, the Grievance Redressal page and the AI Content Disclosure. If they conflict, this order applies: (a) a separate written agreement signed by us; (b) the AUP and the Voice Clone Consent; (c) these Terms; (d) the other policies. [LAWYER: confirm precedence order]

1.4 You accept these Terms when you tick the box (or tap "Continue") at sign-up, or by using the Service. These Terms are an electronic contract under section 10A of the Information Technology Act, 2000. We record which version you accepted and when.

1.5 If you use the Service on behalf of a company or other organisation, you confirm you are authorised to accept these Terms for it, and "you" includes that organisation.

2. Definitions

Swipe sideways to see the whole table →

3. Eligibility and your account

3.1 Age. You must be at least 18 years old. You confirm this when you sign up. We do not knowingly allow anyone under 18 to use the Service, and we will close an account if we learn its user is under 18.

3.2 India. The Service is designed for and offered to users in India. [OWNER: decide whether to block sign-ups from outside India — see README] [LAWYER: effect on DPDP territorial scope and on the Gemini API "available regions" terms]

3.3 Sign-in. You sign in with Google or with a one-time code sent to your email. There are no passwords. Keep access to your email and Google account secure. You are responsible for activity in your account; tell us at once at grievance@{{DOMAIN}} if you think someone else has used it.

3.4 Accurate information, one account. Give accurate information. Do not create multiple accounts to obtain extra free Credits or to get around limits or a suspension.

3.5 Workspaces. Credits, content and Voice Clones belong to the Workspace. Where team features exist, the Workspace owner controls who is a member and what they can do.

4. Public beta, Model 1 and the non-commercial restriction

4.1 Beta. The Service is in public beta. Features may change, be limited or be removed. Output quality varies, especially for languages marked "Beta". Capacity is limited: speech is generated on a single GPU server operated by us in India, so jobs may queue or the Service may be temporarily unavailable.

4.2 Model 1. The beta uses Model 1, which relies on third-party pre-trained model weights licensed under a Creative Commons Attribution-NonCommercial licence. [VERIFY: licence version — the upstream model card says only "CC-BY-NC"] Model 1 also includes an audio tokenizer built with Higgs Materials licensed from Boson AI USA, Inc. under the Boson Higgs Audio 2 Community License, which is based on the Meta Llama 3 Community License. We are not affiliated with, or endorsed by, the authors or licensors of these components. Attributions, the required notices and the licence texts are on our credits page. Your use of Model 1 must also respect the Meta Llama 3 Acceptable Use Policy, which our Acceptable Use Policy includes.

4.3 Non-commercial use only. While Model 1 is used, you may use the Service and its Outputs only for Non-Commercial Purposes. No plan, coupon, redeem code or other arrangement grants commercial rights during the Beta.

Swipe sideways to see the whole table →

4.4 Beta Outputs stay non-commercial. Outputs generated with Model 1 remain limited to Non-Commercial Purposes even after paid launch, unless you generate them again with Model 2 under a plan that includes commercial use. [LAWYER: whether a non-commercial licence on model weights legally restricts Outputs is unsettled; this is a deliberately conservative position]

4.5 The Beta is free. We do not charge for the Beta. While Model 1 is the default model there are no paid plans, top-ups or sold redeem codes. Some invited testers may test the payment flow in Razorpay test mode, where no real money moves.

4.6 Paid launch. When Model 2 launches we will publish updated Terms, plans and commercial-use scopes and give you at least {{PAID_LAUNCH_NOTICE_DAYS}} days' notice. [OWNER: notice period, e.g. 30 days]

5. Your content and Outputs

5.1 You keep your rights in Input. You confirm that you own your Input or have every right, licence and permission needed to submit it — including copyright in text you convert, and the consent of any person whose voice appears in audio or video you upload for transcription or dubbing.

5.2 Licence to us. You give us a non-exclusive, royalty-free licence to host, copy, process, transmit and display your Input and Output, only as needed to (a) provide the Service to you, (b) keep it secure, prevent abuse and comply with law, and (c) keep the provenance records described in §10. The licence ends when the content is deleted, except for copies we must keep under the Privacy Policy.

5.3 No model training on your data. We do not use your Input, Output or voice recordings to train or fine-tune machine-learning models. Audio generated with Model 1 is never used to train any model. If we ever offer a programme where you can choose to contribute your own recordings, it will need your separate, explicit opt-in (off by default, withdrawable at any time) and an updated version of these Terms and our Privacy Policy. Counting credits, fixing errors and measuring service health are not training.

5.4 Outputs. As between you and us, and subject to these Terms (including §4 and §5.5), we do not claim ownership of the Outputs you generate, and you may use them within the scope these Terms allow. We do not promise that an Output is unique, that similar Output will not be generated for others, or that it is protected by copyright. [LAWYER: copyright status of AI-generated audio under the Copyright Act, 1957 — "computer-generated work", s.2(d)(vi) — is uncertain]

5.5 Library Voices. Using a Library Voice lets you use the resulting Output as these Terms allow. You get no rights in the voice itself, the character's name or persona, or the identity of the voice actor. You must not say or suggest that an Output is the voice actor speaking, and you must not use Outputs or previews to clone, recreate or extract a Library Voice.

5.6 Your responsibility. You are responsible for your Input and for how you use Outputs. We use automated and human moderation (§11) but we do not review everything.

5.7 Keep your own copies. Uploads are deleted a few days after your job finishes, and Outputs are kept only for your plan's history period (shown on the Pricing page). Download what you need. Details: Privacy Policy.

6. Voice cloning

6.1 Only your own voice. You may create a Voice Clone only from recordings of your own voice, only if you are 18 or older, and only after giving the Voice Clone Consent and passing the live challenge-sentence check. We may review clones manually and may refuse any clone.

6.2 Private. A Voice Clone is private to your Workspace. We never add it to the public library or use it for anyone else.

6.3 Withdraw any time. You can delete a Voice Clone at any time, which also withdraws your consent. The effects are explained in the Voice Clone Consent.

6.4 Checks. We compare voice samples against our Library Voices and a list of well-known Indian public figures, and we may block, disable or review a clone that appears to belong to someone else.

6.5 Speakers in your uploads. Cloning the original speaker in a dubbing project is off by default and needs verification that you are that speaker, or our manual review.

7. Acceptable use

7.1 You must follow the Acceptable Use Policy. In particular, you must not use the Service to impersonate any person without consent, cheat or defraud anyone, make political deepfakes or election content that misrepresents people, harm or sexualise children, create sexual content, spread hate, or break any law.

7.2 You must not interfere with or abuse the Service — for example by scraping it, bypassing limits, reverse-engineering it, or using Outputs or the Service to build or train a competing speech model.

8. Credits, plans and payments

8.1 Credits are not money. Credits measure use of the Service. They have no cash value, are not legal tender, earn no interest, cannot be transferred, sold or exchanged, and are refundable only as the Refund and Cancellation Policy says. Credits belong to a Workspace.

8.2 How Credits are counted. Text-to-speech uses one Credit per visible character (grapheme) of your original text; emotion tags, pause tags and speaker labels are free. Audio features are counted per minute of audio, rounded up to the next 15 seconds. Current rates are shown on the Pricing page and in your Usage page. We may change rates for future jobs; jobs already running finish at the old rate.

8.3 Reservation. When a job starts we reserve the estimated Credits; when it finishes we charge the actual amount and release the rest. Credits for failed or cancelled work are returned.

8.4 Free Credits. The Free plan includes a monthly Credit allowance and daily limits, shown on the Pricing page. Free Credits are granted each month from your sign-up date (after you verify your email), expire at the end of each monthly period and do not roll over. The monthly grant may pause if you do not sign in for a long period and restarts when you sign in again.

8.5 Promotional Credits and codes. Bonus Credits, coupons and redeem codes have no monetary value, may expire, are limited to one per person unless stated otherwise, and may be withdrawn if obtained through abuse. During the Beta, codes are given only by us and must not be sold.

8.6 Paid plans (only after paid launch). When paid plans exist: payments are processed by Razorpay and we never receive your full card number or UPI PIN; prices are shown on the Pricing page, including how GST applies [CA: GST registration status and whether list prices are shown inclusive or exclusive of GST]; subscriptions renew automatically until cancelled; cancellation takes effect at the end of the current billing period; if a renewal payment fails we give a short grace period before the Workspace moves to the Free plan; upgrades take effect immediately and downgrades at the end of the billing period. We issue invoices as required by law.

8.7 Refunds are covered by the Refund and Cancellation Policy.

8.8 Abuse. Credits obtained by fraud, multiple accounts or payment abuse may be cancelled. If you raise a chargeback or payment dispute, the related Credits may be frozen until it is resolved — please contact us first.

9. API terms

9.1 Access. API access is available only on plans that include it. During the Beta it is limited to testers we approve, and only for Non-Commercial Purposes (§4).

9.2 API keys. Keep your keys secret and never put them in public code or client apps. You are responsible for all use of your keys. We store only a hash of each key and may revoke a key that appears to be exposed or misused. You can revoke and rotate keys at any time.

9.3 Limits. Rate limits, concurrent-job limits, request-size limits and job time limits apply by plan and are documented in the API reference. Respect 429 and 503 responses and the Retry-After header. Do not try to get around limits (for example with multiple accounts or keys).

9.4 Model numbers. You choose a model with the model number. Model 1 is licensed for non-commercial use only (the GET /v1/models response says so). Before we retire a model number we will give at least 90 days' notice by email and the Deprecation response header.

9.5 Webhooks. Webhook endpoints must use HTTPS. We sign webhook requests; verify the signature. We may pause an endpoint that keeps failing.

9.6 If you build on the API. You are responsible for your application and its users. You must (a) make sure your users follow the AUP; (b) tell them clearly that the audio is AI-generated and keep our provenance metadata intact; (c) give them your own privacy notice; and (d) never let anyone clone a voice except the speaker themselves through our verification flow. Reselling, white-labelling or sub-licensing the Service needs our written agreement. [LAWYER: after paid launch, business customers will need a data processing agreement and B2B terms]

9.7 Not allowed. Load-testing or security-testing without our written permission; scraping voices, previews or the voice catalogue; using the API to build a competing service or dataset.

9.8 Changes. The API is versioned (/v1). We will try to keep it backwards-compatible and announce breaking changes in advance, but Beta features may change without notice.

10. AI-generated content: labels and provenance

10.1 All Output is AI-generated. Audio made with DamruVani is synthetically generated information.

10.2 Labels we apply. Every audio file carries embedded metadata identifying it as AI-generated by DamruVani, with a unique job identifier, and we keep a record linking each file's fingerprint to the job that made it. Free-plan audio also carries an audible tag [OWNER: keep only if the Free audible tag is live at publish]. We may add an audible "AI-generated" disclosure at the start of audio and other markers where the law requires it. Details: AI Content Disclosure.

10.3 Do not remove labels. You must not remove, alter, hide or suppress any label, metadata, watermark or other provenance marker, or use any tool to do so.

10.4 Disclose when you publish. When you publish or share Output, you must disclose that it is AI-generated wherever the law or the platform requires it — for example the synthetic-content declarations that significant social media platforms must ask for under the IT Rules, and the Election Commission of India's labelling advisories.

10.5 Do not mislead. Do not present Output as a genuine recording of a real person, or of a real event, in a way that is likely to deceive.

11. Moderation, suspension and termination

11.1 Moderation. We use automated tools (including AI-based text moderation) and human review to prevent prohibited content. A generation may be refused automatically. Requests involving cloned voices are checked more strictly.

11.2 Our actions. If we believe you have broken these Terms, the AUP or the law, or if the law requires it, we may — with or without notice, as the situation requires — refuse or stop a job; remove or disable access to content; disable a Voice Clone or API key; suspend or terminate your account; report the matter to the appropriate authority where the law requires reporting; and, where the law allows, disclose your identity to a complainant who is a victim of the violation (see §12).

11.3 Appeals. You can challenge a decision through Grievance Redressal. Where the IT Rules apply, you may also appeal a Grievance Officer's decision to the Grievance Appellate Committee.

11.4 Closing your account. You can delete your account at any time from Settings → Privacy, or at /account/delete. Deletion starts after a 7-day grace period, during which you can cancel it.

11.5 Effect of termination. When your account ends, your right to use the Service ends, unused Credits lapse (subject to the Refund Policy and to any law that requires otherwise), and your data is deleted as the Privacy Policy describes, except records we must keep by law.

11.6 Ending the Beta or the Service. We may end the Beta or stop offering the Service or any feature. Where practical we will give at least 30 days' notice so you can download your Outputs.

12. Notice to users under the IT Rules

As required by Rule 3(1)(c) and (ca) of the IT Rules, as amended in 2026 [VERIFY: confirm rule numbering against the Gazette text of G.S.R. 120(E)], we inform you — here and by notice at least once every three months — that:

1. If you do not comply with these Terms, the Privacy Policy or the AUP, we have the right to terminate or suspend your access or usage rights immediately, and/or to remove or disable access to the non-compliant information.
2. If your non-compliance involves creating, generating, modifying, hosting, uploading, publishing, sharing or otherwise disseminating information in breach of any law, you may be liable to penalty or punishment under the Information Technology Act, 2000 or other applicable law.
3. If the violation is an offence that must be reported by law (for example under the Bharatiya Nagarik Suraksha Sanhita, 2023 or the Protection of Children from Sexual Offences Act, 2012), we will report it to the appropriate authority as the law requires.
4. Because the Service lets you create synthetically generated information: directing or using the Service to create, modify, publish or share synthetically generated information in breach of the law may attract penalty or punishment under applicable law, including the Information Technology Act, 2000, the Bharatiya Nyaya Sanhita, 2023, the Protection of Children from Sexual Offences Act, 2012, the Representation of the People Act, 1951, the Indecent Representation of Women (Prohibition) Act, 1986, the Sexual Harassment of Women at Workplace (Prevention, Prohibition and Redressal) Act, 2013 and the Immoral Traffic (Prevention) Act, 1956.
5. Such a violation may lead to: (I) immediate disabling of access to or removal of the information; (II) suspension or termination of the violating user's account, without destroying evidence; (III) in accordance with applicable law, identification of the violating user and disclosure of their identity to a complainant who is a victim (or acting on behalf of a victim); and (IV) reporting to the appropriate authority where the violation is an offence that must be reported.

13. Our intellectual property

13.1 The Service, its software, design, the DamruVani name and logo, the Library Voices, voice previews and documentation belong to us or our licensors. [LAWYER: trademark filing for "DamruVani" / डमरूवाणी has not been made — see README] Except for the rights these Terms give you, you get no rights in them.

13.2 You must not copy, modify, distribute, sell or lease any part of the Service, or reverse-engineer or extract its source code, models or voices, except where the law expressly permits it.

13.3 If you send us feedback or ideas, we may use them without any obligation to you.

14. Third-party services and attribution

14.1 Parts of the Service rely on third parties — for example Google (sign-in and, for some text features, the Gemini API), Razorpay (payments), Cloudflare (network security) and email delivery providers. Their own terms and privacy policies also apply when you use their part of the flow. The Privacy Policy lists who processes your data.

14.2 Model 1 uses third-party model weights (OmniVoice, CC BY-NC), an audio tokenizer built with Higgs Materials licensed from Boson AI USA, Inc. and Meta Llama 3 licensed under the Meta Llama 3 Community License, and open-source software. The required notices, links to and copies of the Boson Higgs Audio 2 Community License Agreement and the Meta Llama 3 Community License Agreement, and all other attributions are on our credits page. [LAWYER: exact notice wording — credits.md §2]

15. Availability and support

15.1 The Service is provided on a best-effort basis. There is no service-level agreement and no guaranteed uptime, processing time or queue position. Speech generation depends on our GPU server in India and may be paused for maintenance, power or network outages; queued jobs wait or expire as described in the documentation, and Credits for expired jobs are returned.

15.2 Support is by email at support@{{DOMAIN}} on a best-effort basis.

16. Disclaimers

16.1 To the extent the law allows, the Service and all Outputs are provided "as is" and "as available", without warranties of any kind, whether express or implied, including fitness for a particular purpose, accuracy and non-infringement.

16.2 AI Output can be wrong: words may be mispronounced, the emotion may not match, numbers may be read incorrectly, and transcripts may contain errors. Check Output before you rely on it or publish it. Do not use the Service where an error could cause injury, financial loss or legal harm (for example emergency announcements or certified legal transcription).

17. Limitation of liability

17.1 To the maximum extent permitted by law, we are not liable for any indirect, incidental, special, consequential or punitive loss, or for loss of profits, revenue, goodwill or data, arising out of or in connection with the Service.

17.2 To the maximum extent permitted by law, our total liability for all claims arising out of or in connection with the Service in any 12-month period is limited to the greater of (a) the amount you paid us for the Service in the 12 months before the event giving rise to the claim, and (b) ₹{{LIABILITY_FLOOR_INR}}. [OWNER: floor amount] [LAWYER: enforceability against consumers under the Consumer Protection Act, 2019]

17.3 Nothing in these Terms limits liability that cannot be limited by law, including liability for fraud.

18. Indemnity

You will indemnify and hold harmless Trinetra AI Labs Private Limited, its directors and personnel against claims, losses, penalties and reasonable legal costs arising from (a) your Input, (b) your use of Outputs, (c) your breach of these Terms or the AUP, or (d) your violation of any law or of anyone's rights, including voice, personality, privacy or intellectual-property rights. [LAWYER: scope and enforceability against individual consumers]

19. Changes to the Service and to these Terms

19.1 We may change the Service at any time, especially during the Beta.

19.2 We may update these Terms. For material changes we will notify you by email or in the app at least {{TERMS_CHANGE_NOTICE_DAYS}} days before they take effect [OWNER: e.g. 15 or 30 days], unless the change is required sooner by law or to address abuse. If you continue to use the Service after the effective date, the updated Terms apply. If you do not agree, stop using the Service and delete your account. Changes that need your fresh consent under data-protection law will be asked for separately.

19.3 Earlier versions are archived at /legal/terms/versions.

20. Governing law and disputes

20.1 These Terms are governed by the laws of India.

20.2 Subject to §20.3, the courts at {{CITY}}, Gujarat have exclusive jurisdiction. [LAWYER: choose city based on the registered office; consider arbitration (Arbitration and Conciliation Act, 1996, seat {{CITY}}) for business customers only]

20.3 Nothing in these Terms takes away any right you have as a consumer to approach a consumer commission or other forum under the Consumer Protection Act, 2019, or the grievance and appeal mechanisms under the IT Rules and the Digital Personal Data Protection Act, 2023.

20.4 Before going to court, please contact our Grievance Officer — most issues can be resolved quickly.

21. Contact and grievances

• Grievance Officer: {{GRIEVANCE_OFFICER_NAME}}, {{GRIEVANCE_OFFICER_DESIGNATION}} — grievance@{{DOMAIN}} — details and timelines at /legal/grievance.
• Support: support@{{DOMAIN}}
• Orders from courts, government agencies and law enforcement: legal@{{DOMAIN}}
• Post: Trinetra AI Labs Private Limited, {{REGISTERED_OFFICE_ADDRESS}}

22. General

22.1 Entire agreement. These Terms (with the policies in §1.3) are the entire agreement between you and us about the Service.

22.2 Severability. If any provision is held invalid or unenforceable, the rest stays in effect and the invalid part is applied to the extent allowed.

22.3 No waiver. If we do not enforce a provision, that is not a waiver.

22.4 Assignment. You may not transfer these Terms without our consent. We may transfer them in connection with a merger, acquisition or sale of assets, with notice to you. [LAWYER: consumer-contract assignment and data-transfer implications]

22.5 Force majeure. We are not responsible for delays or failures caused by events beyond our reasonable control, including power or internet outages, natural disasters, government action or failures of third-party providers.

22.6 Language. These Terms are written in English. If we provide a translation, the English version prevails to the extent of any conflict. [LAWYER: confirm; consider Gujarati and Hindi translations for consumers]

22.7 Notices. We may send notices to the email address on your account or show them in the Service. You can send notices to legal@{{DOMAIN}} or by post to our registered office.