Public offer

On providing access to the NCSpeech Studio speech recognition and synthesis platform (studio.ncspeech.ai)

Almaty. Edition of 15 October 2026.

This is a translation of the Russian original. If the versions differ, the Russian text prevails.

NCSPEECH Limited Liability Partnership (LLP “NCSPEECH”), BIN 250340013230, represented by Director Tseren Vladimirovich Andzhukaev, acting under the Charter (the “Provider”), publishes this public offer (the “Offer”) and invites any person to enter into an agreement for access to the NCSpeech Studio platform on the terms set out below.

Under clause 5 of Article 395 of the Civil Code of the Republic of Kazakhstan this Offer is public. Full and unconditional acceptance (acceptance) of the Offer under clause 3 of Article 396 of the Civil Code of the Republic of Kazakhstan is registration of an account on the Platform or payment of a Plan. From the moment of acceptance the agreement between the Provider and the User is deemed concluded in writing (the “Agreement”).

Read the Offer carefully before signing up. If you disagree with any of its terms, do not register on the Platform and do not pay for a Plan.

1. Terms and definitions

1.1. The following terms are used in the Offer:

• a) Platform: the Provider’s hardware-and-software complex “NCSpeech Studio” at https://studio.ncspeech.ai, including speech-recognition (ASR) and speech-synthesis (TTS) models for Kazakh and Russian, a web interface and an application programming interface (API).
• b) User: a natural person aged 18 or over, an individual entrepreneur or a legal entity that has accepted the Offer. A person who registers an account on behalf of a legal entity confirms that they have the necessary authority.
• c) Account (Cabinet): a protected section of the Platform available to the User after registration and authentication, in which the User manages the Plan, API keys, voices and settings and sees credit usage, payment history and recognition history.
• d) API key: a unique identifier created by the User in the Cabinet to authenticate API requests.
• e) Credit: a unit of accounting for the volume of Services. One second of recognition of one audio channel equals 1 credit; one second of streaming recognition of one audio channel equals 1.25 credits; one second of synthesized speech equals 12 credits. A stereo recording counts as two channels. Creating one user voice deducts 100 credits.
• f) Plan: the set of terms on which Services are provided (price, credit volume for a Billing period, number of user voices, number of Concurrent streams, set of functions), described in section 4 of the Offer and on the “Plans” page of the Platform.
• g) Subscription: a paid Plan with monthly automatic renewal and charging of its price to the User’s bank card.
• h) Billing period: one month counted from the date the Plan is paid. Each next Billing period starts on the same calendar day of the next month, or on the last day of the month if that day does not exist.
• i) Concurrent stream: one User request that the Platform is processing at a given moment. A streaming recognition session holds one Concurrent stream for its whole duration. The number of simultaneously processed requests is limited by the Plan.
• j) User data: audio recordings, texts, word lists, pronunciation dictionaries, voice samples and other data that the User submits to the Platform.
• k) Results: text obtained when recognising User data, synthesized speech and related data (timestamps, speaker labels, confidence scores) that the Platform returns to the User.
• l) Documentation: a description of API methods, data formats, technical limits and rules for using the Platform, published at https://studio.ncspeech.ai/docs.
• m) Payment service: the FreedomPay and Robokassa internet acquiring services through which the User pays for a Plan by bank card.
• n) Privacy policy: the Provider’s document on processing of Users’ personal data, published on the Platform and an integral part of the Agreement.

2. Subject of the agreement

2.1. The Provider grants the User remote access to the Platform through the web interface and API for speech recognition and synthesis in the volume and on the terms of the chosen Plan (the “Services”), and the User accepts and pays for the Services in the manner set by the Offer.

2.2. Access to the Platform is granted on a subscription model: during each Billing period the User has the credit volume, number of user voices and number of Concurrent streams set by the Plan.

2.3. The Provider grants the User the right to use the Platform to the extent needed to receive the Services, on a simple (non-exclusive) licence for the term of the Agreement, without the right of assignment or sublicensing. Exclusive rights in the Platform, models, voices, Documentation and the Provider’s software are not transferred to the User.

2.4. The User may use the Results without territorial or manner-of-use limits, including embedding them in the User’s own products and services and providing them to the User’s clients, subject to the limits in clause 7.4 of the Offer.

2.5. The Services do not include hosting models in the User’s infrastructure, further training of models or configuring voice assistants. Such work is provided under a separate agreement whose terms are agreed on request to info@ncspeech.org.

2.6. The Services are provided remotely, in electronic form. No physical goods are delivered under the Agreement.

3. Registration and account

3.1. To obtain access to the Platform the User registers an Account, stating a working e-mail address and signing in through the Platform’s authentication service (e-mail and password or a Google account). Registration means acceptance of the Offer, agreement with the Privacy policy and confirmation that the User is 18 or older.

3.2. The User undertakes to give accurate details at registration and keep them current. The e-mail stated at registration is used for sign-in, notices of payment and Subscription renewal, and for legally significant communications under the Agreement.

3.3. The User keeps sign-in credentials and API keys confidential. All actions performed using the User’s Account or API keys are deemed performed by the User. If compromise is suspected, the User immediately changes the password, revokes the API key in the Cabinet and notifies the Provider.

3.4. The User may delete the Account at any time in the “Settings” section of the Cabinet (with confirmation of the Account e-mail) or by sending a request to info@ncspeech.org from that address. Deleting the Account terminates the Agreement under section 12 of the Offer.

4. Plans, credits and subscription

4.1. The Plans in the table below apply on the Platform. All prices are in tenge; the services are not subject to VAT (clause 5.3 of the Offer). Hours are a planning guide calculated from credit-deduction rates; accounting is in credits.

4.2. Credit-deduction rates: speech recognition 1 credit per second of audio per channel (3,600 credits per hour of mono, 7,200 credits per hour of stereo); streaming recognition (audio sent while it is recorded, with text returned during the session) 1.25 credits per second of sent audio per channel (4,500 credits per hour of mono); speech synthesis 12 credits per second of synthesized speech (720 credits per minute); creating a user voice from a sample 100 credits. Credits are shared between recognition and synthesis.

4.3. The Free Plan is granted at registration without linking a bank card. The starting credit quota is granted once for one Billing period and is not topped up. When that period ends, processing of new requests is suspended until the User pays for a paid Plan. The Free Plan is not a trial of a paid Subscription: there is no automatic switch to a paid Plan, and the first charge to a card is possible only after the User has chosen a paid Plan and paid for it.

4.4. A paid Plan is taken out as a Subscription. At the start of each Billing period the User is granted the Plan’s credit volume. Unused credits expire at the end of the Billing period and are not carried to the next period; their value is not refunded except in cases expressly provided in section 6 of the Offer.

4.5. When Billing-period credits are exhausted, processing of new requests is suspended until the next Billing period starts, extra credits are bought or the User switches to a higher Plan. No charge to the card above the Plan price is made without a separate action by the User. A request accepted for processing is completed and paid in full even if the credit balance becomes negative as a result; while the balance is negative, new requests are not accepted.

4.6. The User may buy extra credits above the Plan volume while the current Billing period of a paid Plan is in force. The User enters an amount in tenge on the “Plans” page within the limits set by the Platform; the credit count is calculated at the current Plan rate (Plan price divided by its credit volume). The top-up is paid as a one-off payment and does not renew automatically. Bought credits are counted in the same pool as Plan credits and expire at the end of the Billing period and when the User switches to another Plan.

4.7. The User may switch to another paid Plan on any day: from the date the new Plan is paid a new Billing period starts with the new Plan’s quota, and unused remainder of the previous Plan’s credits expires and is not carried to the new period.

4.8. The number of Concurrent streams is limited by the Plan. Requests above the limit are not queued and are rejected; the API returns status 429 and the User retries later. Rejection above the limit is not unavailability of the Platform.

4.9. The volume of Services provided is determined from the Platform’s accounting data shown in the Cabinet (number of requests, duration of processed audio and synthesized speech, credits deducted). If the User has reasoned objections, the Provider reconciles the accounting data within 5 (five) business days.

4.10. Technical limits of the Platform (supported audio formats, maximum file size and text length, request-rate limits) are set by the Documentation. The Provider may change them if that does not reduce the volume of Services under the Plan.

4.11. A streaming recognition session lasts at most 30 (thirty) minutes of sent audio, after which the Platform finishes recognizing the sent audio and closes the session. Credits are deducted as audio is sent and for the remainder when the session ends. A started session continues until it ends even if the credit balance becomes negative. The session result is kept in the recognition history.

5. Price of services and payment procedure

5.1. Plan prices are stated in section 4 of the Offer and on the “Plans” page of the Platform in tenge. All settlements under the Agreement are in tenge.

5.2. Payment is made by a Visa or Mastercard bank card through the Payment service. Card details are entered by the User on a protected page of the Payment service; the Provider does not receive or store the full card number, expiry date or security code (masked number — clause 5.5a). The payment is confirmed by the issuing bank’s 3-D Secure technology.

5.3. The Provider is a participant of the Astana Hub international technopark and is not registered for value-added tax; the price of the Services is not subject to VAT.

5.4. The Services are prepaid for each Billing period. The payment date is the date the Payment service successfully confirms the payment. Access under the paid Plan is activated automatically immediately after payment confirmation, and confirmation of the payment is sent to the User at the Account e-mail.

5.5. By taking out a Subscription the User instructs the Provider and the Payment service, monthly, on the start date of each next Billing period, to charge the chosen Plan price to the stated bank card without further confirmation (a recurring payment). In doing so:

• a) card details are stored by the Payment service as a protected token; the Provider does not receive or store the full card number, expiry date or security code; the Provider stores only a masked card number (last four digits) for display in the Cabinet;
• b) the User may at any time refuse automatic renewal by cancelling the Subscription in the “Plans” section of the Cabinet, and unlink the card in “Settings” or by sending a request to info@ncspeech.org;
• c) if the charge fails (insufficient funds, blocked card or expired card), the Provider retries; if payment still does not arrive, automatic renewal stops. Access for the already paid Billing period continues until it ends. When the period ends without a successful payment, processing of new requests is suspended until a Plan is paid for again. No debt arises.

5.6. For Users that are legal entities or individual entrepreneurs, on request to info@ncspeech.org the Provider sends an electronic statement of services for the Billing period. The statement is based on the Platform’s accounting data.

5.7. The Provider may change Plan prices and Plan contents. The new price is published on the Platform at least 30 (thirty) calendar days before it applies and applies to the User’s Subscription only from the start of the next Billing period. If the User disagrees with the new price, the User may cancel the Subscription before that Billing period starts.

6. Cancelling the subscription and refunds

6.1. The User may cancel the Subscription at any time in the “Plans” section of the Cabinet or by sending a request to info@ncspeech.org. After cancellation, access under the paid Plan and unused credits (including bought credits) remain until the end of the current Billing period; further charges are not made. When the period ends, processing of new requests is suspended until a paid Plan is paid for. The Account, API keys, voices and settings are kept. There is no automatic switch to the Free Plan and no re-grant of its quota.

6.2. The Provider refunds money paid by the User in the following cases:

• a) the User withdraws from a paid Plan within 14 (fourteen) calendar days of the payment date and no credits of the paid Billing period have been used: the full price of the Billing period is refunded;
• b) the Services were not provided through the Provider’s fault: the Platform was unavailable for more than 48 (forty-eight) hours in a row, or the Platform could not process valid User data, as confirmed by the Provider’s technical support. Part of the Billing-period price in proportion to the downtime is refunded or, at the User’s choice, compensation is granted in credits;
• c) a payment was charged in error (a duplicate charge, a charge after cancellation of the Subscription, a charge of an amount that does not match the Plan): the amount of the erroneous payment is refunded.

6.3. In other cases the price of paid Services is not refunded, because the Services are provided automatically and are consumed as credits are used. The price of a partly used Billing period, the price of bought credits after they have been used and the price of unused credits at the end of the Billing period are not refundable.

6.4. To obtain a refund the User sends a request to info@ncspeech.org from the Account e-mail, stating the date and amount of the payment and the reason. The Provider reviews the request within 5 (five) business days and notifies the User of the decision.

6.5. The refund is made to the same bank card the payment came from, through the Payment service. The Provider initiates the refund within 5 (five) business days of the decision. The time for the funds to reach the card is set by the issuing bank and is usually 3 (three) to 10 (ten) banking days. A cash refund or a refund to another card is not made.

6.6. If the full price of the Billing period is refunded, access under the paid Plan ends from the date of the refund decision. Processing of new requests is suspended until a paid Plan is paid for. There is no automatic switch to the Free Plan and no re-grant of its quota.

7. Rights and obligations of the parties

7.1. The Provider undertakes to:

• a) grant the User access to the Platform and provide the Services in accordance with the Offer, the Plan and the Documentation;
• b) process User data solely to provide the Services on the terms of section 9 of the Offer and the Privacy policy and not disclose them to third parties except where expressly provided by the Agreement or the legislation of the Republic of Kazakhstan;
• c) notify the User of planned work affecting availability of the Platform at least 24 (twenty-four) hours in advance at the Account e-mail or in the Platform interface; urgent work to remove security threats may be carried out without prior notice;
• d) provide technical support on operation of the Platform, payment, the Subscription and refunds at info@ncspeech.org and by phone +7 (771) 826-03-23 on business days from 09:00 to 18:00 Astana time; a reply is sent within 1 (one) business day.

7.2. The Provider may:

• a) update the Platform, models, voices and Documentation. Of API changes that break backward compatibility the Provider notifies Users at least 30 (thirty) calendar days in advance;
• b) introduce technical limits needed to protect the Platform and its other users from overload and abuse, if such limits do not reduce the volume of Services under the Plan;
• c) suspend access to the Platform in the cases provided in section 12 of the Offer;
• d) send the User notices related to operation of the Platform, payment and changes to the Offer at the Account e-mail.

7.3. The User undertakes to:

• a) pay for the Services in the manner and within the time set by the Offer;
• b) use the Platform in accordance with the Offer, the Documentation and the legislation of the Republic of Kazakhstan;
• c) have all rights and permissions needed to submit User data to the Platform and have them processed, including consents of personal-data subjects, and when using voice samples to create a user voice obtain the consent of the person whose voice is used;
• d) not submit to the Platform data whose processing by the Provider is prohibited by law, including information constituting state secrets;
• e) keep sign-in credentials and API keys confidential.

7.4. The User is prohibited from:

• a) disclosing sign-in credentials and API keys to third parties, and granting third parties direct access to the Platform API;
• b) reselling, leasing or otherwise providing third parties with access to the Platform as a standalone speech-recognition or speech-synthesis service. Using the Results in the User’s own products and services, including for the User’s clients, is not a breach of this clause;
• c) researching, decompiling, extracting or reproducing models, weights and other components of the Platform, and using the Results to train models that compete with the Platform;
• d) circumventing technical limits of the Platform, conducting load testing without the Provider’s written agreement, interfering with the Platform or attempting unauthorised access to its components and other users’ data;
• e) using the Platform to create materials that mislead as to the identity of a speaker, to imitate a real person’s voice without their consent, for fraud, for distributing prohibited information and for other unlawful purposes.

8. Intellectual property and data

8.1. The Platform, models, voices (other than user voices created by the User from the User’s own samples), Documentation, trade marks and other intellectual-property objects of the Provider belong to the Provider. The Agreement does not grant the User any rights in them other than those expressly stated in clause 2.3 of the Offer.

8.2. User data belong to the User. The User grants the Provider the right to process User data solely to the extent needed to provide the Services. The Provider does not acquire rights in User data and does not use them to train models, evaluate them or for other own purposes without the User’s separate written consent.

8.3. Rights in the Results belong to the User from the moment they are formed. The Provider does not claim the Results and does not use them for its own purposes. The User acknowledges that the Results are formed automatically and that similar or matching Results may be obtained by other users of the Platform when similar data are processed.

8.4. User voices created in the Cabinet from the User’s own samples are used only in the User’s Account and are not provided to other users of the Platform. The User is responsible for having the consent of the person whose voice was used to create such a voice.

8.5. Results of automatic speech recognition and synthesis may contain errors, including distortion of words, numbers, names, abbreviations and special terms. The User independently decides whether the Results may be used in the User’s processes and ensures human review of the Results where an error may cause harm to life, health, property or third-party rights.

9. Personal data and confidentiality

9.1. The Provider processes the User’s personal data (e-mail address, Account data, information about payments and use of the Platform) in accordance with the Law of the Republic of Kazakhstan dated 21 May 2013 No. 94-V “On personal data and their protection” and the Privacy policy. By registering on the Platform the User consents to the collection and processing of their personal data for the purposes and in the volume stated in the Privacy policy.

9.2. Personal data of third parties contained in User data (voice, content of speech and texts) are processed by the Provider on the User’s instructions and solely to return the Result. The User is the owner and (or) operator of the database containing such data and is responsible for having a lawful basis to submit them to the Platform.

9.3. User data are not retained by the Provider after the Result is returned to the User, except: (a) user voices created in the Cabinet; (b) recognition history in the Cabinet (Result text and related request metadata; the latest 1,000 records, older ones are deleted automatically), which the User may delete at any time; (c) word-hint lists stored together with that history record. The voice sample is not stored separately after the clone is created. The accounting system retains request metadata (date and time, duration, credits deducted, API-key identifier, response code).

9.4. The Provider processes User data on server equipment located in the Republic of Kazakhstan and does not carry out their cross-border transfer.

9.5. The Parties undertake not to disclose to third parties confidential information obtained in connection with the Agreement, including User data, Results, sign-in credentials and API keys, except as provided by the legislation of the Republic of Kazakhstan. Confidentiality obligations last for the term of the Agreement and 3 (three) years after it ends, and indefinitely as to personal data.

9.6. For Users that are legal entities, on request to info@ncspeech.org the Provider provides a personal-data processing agreement describing protection measures and the place of processing.

10. Liability of the parties

10.1. For non-performance or improper performance of obligations under the Agreement the Parties are liable in accordance with the Agreement and the legislation of the Republic of Kazakhstan.

10.2. The Provider’s aggregate liability under the Agreement is limited to the amount paid by the User for the Billing period in which the event giving rise to liability occurred. The limit does not apply to liability for wilful breach, or where limitation of liability is not allowed by the legislation of the Republic of Kazakhstan, including consumer-protection legislation.

10.3. Neither Party compensates the other for lost profits or indirect losses, or for losses caused by inability to use the Platform due to circumstances for which the Party is not responsible.

10.4. The User is solely responsible for the content of User data, the lawfulness of their collection and submission to the Platform, use of the Results and the acts of persons to whom the User granted access to the Account. The User compensates the Provider for documented losses, including fines and third-party claims, caused by the User’s breach of clauses 7.3 and 7.4 of the Offer.

10.5. The Provider is not liable for inaccuracy of the Results if the Platform operated in accordance with the Documentation, for failures of communications networks and equipment not belonging to the Provider, for non-operation of an integration made by the User in breach of the Documentation, for acts of the Payment service and the issuing bank in processing payments, or for consequences of the User using the Results without review in the cases in clause 8.5 of the Offer.

10.6. The Parties are released from liability for non-performance caused by force majeure: natural disasters, fires, military action, mass unrest, epidemics, acts of state authorities, large-scale failures of power and communications networks. If such circumstances last more than 2 (two) months, either Party may terminate the Agreement with settlement for Services actually provided.

11. Changes to the offer

11.1. The Provider may unilaterally change the terms of the Offer. The new edition is published on the Platform with the edition date at least 10 (ten) calendar days before it takes effect. Changes to Plan prices are made under clause 5.7 of the Offer.

11.2. Continued use of the Platform after the new edition takes effect means the User agrees to the changes. If the User disagrees with the new edition, the User may stop using the Platform and delete the Account before the changes take effect; the previous edition applies to a paid and unfinished Billing period.

11.3. The current edition of the Offer is permanently published at https://studio.ncspeech.ai. On the User’s request to info@ncspeech.org the Provider sends a copy of the edition that was in force on the date of acceptance.

12. Term, suspension and termination of the agreement

12.1. The Agreement takes effect from acceptance of the Offer and continues until terminated by one of the Parties in the manner provided in this section.

12.2. The User may terminate the Agreement at any time by cancelling the Subscription (section 6 of the Offer) and deleting the Account. A paid Billing period is not shortened or refunded except in the cases in clause 6.2 of the Offer.

12.3. The Provider may suspend the User’s access to the Platform in whole or in part, notifying the User at the Account e-mail or in the Platform interface, in the following cases:

• a) the User’s breach of clause 7.4 of the Offer or use of the Platform in a way that threatens its operation, security or third-party rights. In this case access may be suspended immediately with notice the same day;
• b) receipt of a binding request of a state authority or a court;
• c) a reasoned suspicion of fraudulent use of a bank card in payment, until the circumstances are clarified with the Payment service.

12.4. Access is restored within 1 (one) business day after the ground for suspension is removed. If the User has not remedied the breach within 10 (ten) business days of the notice, the Provider may terminate the Agreement and delete the Account. The price of the Billing period in which access was suspended through the User’s fault is not refunded.

12.5. The Provider may terminate the Agreement with a User who has no active Subscription if the User has not signed in to the Account for more than 12 (twelve) consecutive months, notifying the User at least 30 (thirty) calendar days in advance at the Account e-mail.

12.6. On termination of the Agreement the Provider revokes API keys, blocks the Account and within 10 (ten) business days deletes Account data, including voice samples and settings, except information about the volume of Services and settlements needed for accounting and dispute resolution. Termination of the Agreement does not release the Parties from obligations that arose before it ended.

13. Dispute resolution and governing law

13.1. Disputes and disagreements under the Agreement are resolved by the Parties through negotiations. A claim is sent to info@ncspeech.org (for the User) or to the Account e-mail (for the Provider); the time to reply to a claim is 10 (ten) business days from receipt.

13.2. If no agreement is reached, the dispute is submitted to a court in accordance with the legislation of the Republic of Kazakhstan. Disputes with Users that are legal entities or individual entrepreneurs are heard by the Specialised Inter-District Economic Court of the city of Almaty. A User who is a consumer may apply to a court in the manner set by the consumer-protection legislation of the Republic of Kazakhstan.

13.3. The Agreement is governed by the legislation of the Republic of Kazakhstan. On matters not regulated by the Offer the Parties are guided by the Civil Code of the Republic of Kazakhstan, the Law of the Republic of Kazakhstan “On protection of consumer rights”, the personal-data law and other acts of legislation of the Republic of Kazakhstan.

13.4. Notices under the Agreement are sent by e-mail: to the User at the Account address, to the Provider at info@ncspeech.org. A message is deemed received on the day it is sent if it is sent on a business day before 18:00 Astana time, and on the next business day in other cases.

14. Provider details