qziraBack to home

Terms of Service

Last updated: February 5, 2026

Version: 3.0

These Terms of Service (“Terms”) govern your use of the AI API gateway service “qzira” (“Service”) provided by 138data (“we,” “our,” or “us”). Please read these Terms carefully before using the Service.

Article 1 (Definitions)

In these Terms, the following terms have the meanings set forth below:

  1. “Service” means the AI API gateway service provided by us under the name “qzira,” including the associated dashboard (app.qzira.com), documentation (docs.qzira.com), website (qzira.com), and all related services.
  2. “User” means any individual or legal entity that agrees to these Terms and registers to use the Service.
  3. “AI Provider” means entities that provide AI model services, including OpenAI, L.L.C., Anthropic, PBC, and Google LLC.
  4. “BYOK Key” means an API key that a User obtains directly from an AI Provider and registers with the Service.
  5. “qzira API Key” means an authentication key issued by us to a User for accessing the Service.
  6. “Agent” means any AI agent, coding assistant (such as Cursor, Claude Code, or Devin), script, automation program, or similar software created, configured, or used by a User.
  7. “Confidential Information” means non-public business, technical, or other information disclosed by one party to the other in connection with the Service, excluding information that: (i) was publicly known at the time of disclosure; (ii) became publicly known after disclosure through no fault of the recipient; (iii) was already in the recipient's possession at the time of disclosure; or (iv) was lawfully obtained from a third party without restriction.
  8. “Antisocial Forces” means organized crime groups, members thereof, quasi-members, affiliated enterprises, corporate extortionists, social movement advocates acting in bad faith, special intelligence violent groups, and similar persons.
  9. “Generated Content” means all content (text, images, code, and any other output) generated through the Service via AI models.

Article 2 (Service Overview)

  1. The Service is a gateway that relays API requests to multiple AI Providers through a unified endpoint.
  2. The Service uses a BYOK (Bring Your Own Key) model. Users use API keys they obtain directly from AI Providers. We do not operate our own AI models; all AI processing is performed on each AI Provider's systems.
  3. The Service provides cost management features such as daily limits, budget alerts, and auto-stop, but these are supplementary tools and their completeness is not guaranteed (see Article 13).

Article 3 (Account Registration)

3-1. Registration Requirements

  1. Users must provide accurate and current information when registering for the Service.
  2. Registration using false information is prohibited.
  3. Creating multiple accounts by a single User is prohibited, including creating multiple accounts to obtain multiple Free tier allocations.
  4. Users are responsible for managing their accounts, including any unauthorized use by third parties.

3-2. API Key Management

  1. Users must not disclose, share, or transfer qzira API Keys or BYOK Keys to third parties.
  2. If a User discovers leakage or unauthorized use of an API Key, they must immediately notify us at https://138io.com/contact and invalidate the affected key.
  3. Charges incurred by AI Providers due to leakage or unauthorized use of API Keys are the User's responsibility, regardless of the cause.

Article 4 (Fees and Payment)

  1. Service fees are as set forth in the plan schedule separately determined by us.
  2. Monthly fees for paid plans are payable by credit card through Stripe.
  3. AI Provider API usage fees are entirely the User's responsibility, regardless of cause. Service fees do not include payments to AI Providers.
  4. We will notify Users at least 30 days in advance of any plan fee changes.
  5. Plan changes (upgrades or downgrades) are limited to once per month.
  6. Usage Measurement: Usage metrics such as request counts and token counts are calculated based on measurement logs in our systems.
  7. Attribution of BYOK Key Charges: Charges incurred with AI Providers through a User's BYOK Key are solely a matter between the User and the AI Provider, regardless of cause, and we have no involvement.
  8. Technical controls provided by us (rate limits, daily caps, monthly caps, auto-stop, etc.) are supplementary tools for usage management. Even if these controls fail to operate correctly, Users acknowledge that they remain obligated to pay charges incurred with AI Providers through their BYOK Keys.

Article 5 (Prohibited Activities)

Users must not engage in any of the following activities when using the Service:

5-1. System Abuse

  1. Agent infinite loops / unlimited requests: Making unlimited consecutive API calls without termination conditions or with inappropriate termination conditions.
  2. Rate limit circumvention: Intentionally circumventing the Service's rate limits by creating multiple accounts or rotating API keys.
  3. Abnormally high request volumes: Concentrating extreme access within a short period, even within daily limits, in a manner that threatens infrastructure stability.
  4. Safety feature circumvention: Attempting to disable or bypass safety features such as budget alerts, auto-stop, or rate limits.

5-2. Security

  1. Unauthorized API key use: Using or sharing another person's BYOK Key or qzira API Key without authorization.
  2. Reverse engineering: Analyzing, imitating, or replicating the Service's systems, APIs, or infrastructure.
  3. Vulnerability scanning: Conducting security testing or vulnerability scanning against the Service's infrastructure without prior written permission.
  4. Prompt injection: Submitting inputs designed to alter or disable the Service's system prompts.

5-3. Content

  1. Illegal content generation: Using AI for illegal purposes such as child exploitation, terrorism, or malware creation.
  2. AI Provider policy violations: Even under the BYOK model, each AI Provider's (OpenAI, Anthropic, Google, etc.) usage policies apply. Submitting prompts that violate these policies is prohibited.
  3. Spam and phishing: Using AI to create mass spam, phishing sites, fraudulent content, or similar materials.

5-4. Account

  1. Reselling or subletting accounts: Reselling, sublicensing, or sharing qzira accounts or API access with third parties.
  2. Fraudulent Free tier acquisition: Creating multiple accounts to obtain multiple Free plan allocations.
  3. Competing service development: Using the Service to develop or sell an AI API gateway service that competes with the Service.
  4. Other misuse: Violating laws or public policy, or any other activities we deem inappropriate.

5-5. Consequences of Violations

If a User violates any of the above prohibited activities, we may take the following actions without prior notice:

  1. Temporary or permanent suspension of API access
  2. Forced account termination
  3. Immediate billing of outstanding charges
  4. Claims for damages (within the limits set forth in Article 16)
  5. Legal action

Article 6 (Agent Use and Responsibility)

  1. Users bear full responsibility for all API requests made and all Generated Content produced by AI agents, scripts, or programs they use, regardless of intent.
  2. Users are responsible for configuring appropriate termination conditions, request limits, and error handling for any agents they use.
  3. The fact that an action was performed autonomously by an Agent does not constitute a defense against charges incurred with AI Providers, damages to third parties, or violations of these Terms arising from such autonomous behavior.

Article 7 (Third-Party AI Providers)

  1. The Service relays API requests using technology provided by third-party AI Providers such as OpenAI, L.L.C., Anthropic, PBC, and Google LLC. We operate as an independent gateway and have no agency, delegation, partnership, or other special relationship with AI Providers.
  2. Under the BYOK model, Users have a direct contractual relationship with each AI Provider. Users are obligated to comply with each AI Provider's terms of service, usage policies, and content policies.
  3. Activities prohibited by AI Providers are also prohibited under the Service.
  4. If the Service's specifications, terms, or prohibited activities are affected by changes to an AI Provider's terms, pricing, or service termination, we will take appropriate measures to the extent possible, but we are not responsible for damages to Users caused by such changes.
  5. We have no involvement in disputes between Users and AI Providers (including billing disputes, content policy violations, or account suspensions).
  6. If we reasonably determine that a User's usage may violate an AI Provider's policies, we reserve the right to suspend that User's API access without prior notice.

Article 8 (Intellectual Property and Data Handling)

  1. Service rights: All intellectual property rights in the Service's software, UI, API design, documentation, and related materials belong to us. Users may use these solely for the purpose of using the Service.
  2. Input data rights: Copyright in text and data submitted to the Service by Users remains with the User or the rightful rights holder.
  3. No training use: We do not use Users' input data (prompts) or AI response data for machine learning. We do not operate AI models, and our architecture makes such use technically impossible. We may use non-personally-identifiable statistical data (such as request counts and token counts) to improve the Service.
  4. Generated Content rights and responsibility: Rights in Generated Content are governed by each AI Provider's terms. Users bear all responsibility for the use, publication, and commercial use of Generated Content.
  5. Prompt responsibility: Users warrant that prompts they submit do not infringe any third party's copyright, trademark, or other intellectual property rights.
  6. No warranty of non-infringement: We do not warrant that Generated Content does not infringe third-party intellectual property rights. AI models are probabilistic language models and may generate output similar to existing works. Users are responsible for confirming that Generated Content does not infringe third-party rights before publishing or using it commercially.
  7. No warranty of safety: We do not warrant the safety, accuracy, or appropriateness of Generated Content. We are not responsible for damages caused by harmful information, inaccurate information, or inappropriate expressions in Generated Content.

Article 9 (Emergency Suspension)

  1. We may temporarily restrict or suspend a User's API access without prior notice (“Emergency Suspension”) if we reasonably determine that any of the following applies:
    • The stable operation of the Service's infrastructure as a whole is being or may be disrupted.
    • Other Users' use of the Service is being or may be significantly impacted.
    • Unauthorized use is suspected.
    • Violation of an AI Provider's policies is suspected.
  2. An Emergency Suspension may be implemented without waiting for a final determination that these Terms have been violated.
  3. If we implement an Emergency Suspension, we will endeavor to notify the User and explain the reasons within a reasonable period.

Article 10 (Service Suspension and Termination)

  1. We may temporarily suspend all or part of the Service without prior notice to Users if any of the following occurs:
    • System maintenance or updates are required.
    • Force majeure events such as earthquakes, lightning, fire, or power outages make operation difficult.
    • Failures or specification changes occur on the AI Provider's side.
    • Failures occur with infrastructure providers such as Cloudflare.
    • We otherwise determine that suspension is necessary.
  2. We may terminate all or part of the Service with 30 days' prior notice, except where unavoidable circumstances require otherwise.
  3. Upon termination of the Service, we will endeavor to refund the pro-rated remaining portion of prepaid monthly fees for the current month, but this is not guaranteed.

Article 11 (Data Handling)

  1. We record and retain metadata necessary for operating the Service (such as request timestamps, AI model names used, token counts, and response times).
  2. The content of prompts submitted by Users and AI responses is only relayed and is not permanently stored on our servers.
  3. Metadata may be retained for a certain period when necessary for audits, issue resolution, or security incident investigations.
  4. The handling of personal information is governed by our separately established Privacy Policy.

Article 12 (Cancellation and Account Deletion)

  1. Users may cancel the Service and request account deletion at any time.
  2. Cancellation of a paid plan does not entitle the User to a pro-rated refund for the remaining days of the current month. The Service remains available until the end of the current month after cancellation.
  3. After an account deletion request, we will completely delete the User's data within 30 days, except for data required to be retained by law (such as billing history).

Article 13 (Disclaimers)

13-1. Limitations of Cost Management Features

  1. Budget alerts and auto-stop features are supplementary tools for cost management. Their complete real-time accuracy and reliability are not guaranteed.
  2. Technical factors (network latency, KV store consistency, email delivery delays, etc.) may cause delays of seconds to minutes in alert notifications and auto-stop.
  3. If API costs to an AI Provider exceed a User's configured limit due to such delays or malfunctions, we are not responsible for the excess amount.
  4. Users acknowledge that ultimate responsibility for cost management lies with them, and we recommend regularly checking usage through AI Provider dashboards as well.

13-2. Service Availability

  1. We may set availability targets for the Service, but no SLA (Service Level Agreement) guarantees are provided.
  2. We are not responsible for service interruptions or quality degradation caused by AI Provider failures, specification changes, or pricing changes.

13-3. AI Output

  1. The Service is a gateway that relays requests to AI Providers. We bear no responsibility whatsoever for the accuracy, appropriateness, or completeness of AI-generated content.
  2. AI models generate probabilistic responses and may produce factually incorrect information (hallucinations), incomplete information, or harmful information. We are not responsible for such output.
  3. We are not responsible for damages resulting from decisions made in reliance on Generated Content (medical, legal, investment decisions, etc.).
  4. We are not responsible for copyright infringement, privacy violations, or other rights violations contained in AI Provider output.

Article 14 (Indemnification)

  1. If damages are caused to us or any third party by any of the following, Users shall indemnify us and compensate us for all damages incurred (including reasonable attorney's fees):
    • User's violation of these Terms
    • Violation of AI Provider policies by the User or their Agents
    • Third-party copyright infringement caused by prompts submitted or Generated Content produced by the User
    • Third-party privacy violations or defamation caused by prompts submitted or Generated Content produced by the User
    • Damages caused by the User's improper management or leakage of API keys
  2. The indemnification obligation in the preceding paragraph applies to all Users regardless of their plan (including Free plan users).

Article 15 (Confidentiality)

  1. Users must not use our Confidential Information learned through use of the Service for any purpose other than using the Service, and must not disclose or leak it to third parties.
  2. We will not use Users' BYOK Keys or other Confidential Information for any purpose other than operating the Service, and will not disclose it unnecessarily to third parties.
  3. Confidentiality obligations survive termination of these Terms for a period of 2 years.

Article 16 (Limitation of Liability)

  1. Except in cases of our willful misconduct or gross negligence, the total amount of damages we owe to any User shall not exceed the monthly service fee paid by that User in the month preceding the month in which the damage occurred.
  2. Notwithstanding the preceding paragraph, our liability to Free plan users shall not exceed ¥0 (zero yen), except in cases of our willful misconduct or gross negligence.
  3. We shall not be liable for indirect damages, special damages, consequential damages, or lost profits under any circumstances.
  4. Scale plan users may negotiate the liability cap in this Article individually through a separate written agreement.

Article 17 (Force Majeure)

  1. We are not responsible for service interruptions, delays, or quality degradation caused by natural disasters (earthquakes, typhoons, floods, etc.), war, terrorism, riots, government regulations, power outages, internet connection failures, AI Provider failures, failures of infrastructure providers such as Cloudflare, or other circumstances beyond our reasonable control (“Force Majeure Events”).
  2. If the Service is suspended due to a Force Majeure Event for more than 30 days, Users may cancel the Service. In such case, we will endeavor to provide a pro-rated refund of prepaid fees.

Article 18 (Exclusion of Antisocial Forces)

  1. Users represent and warrant that they do not currently and will not in the future constitute Antisocial Forces or have close relationships with Antisocial Forces.
  2. If we discover that a User has violated the preceding paragraph, we may immediately terminate these Terms and delete that User's account without notice.
  3. We bear no responsibility for any damages suffered by a User due to termination under the preceding paragraph.

Article 19 (Changes to These Terms)

  1. When we change these Terms, we will publish the changes on the Service and notify Users of material changes at least 30 days in advance.
  2. Continued use of the Service after changes take effect constitutes acceptance of the revised Terms.
  3. If a User does not agree to changes to the Terms, they may cancel the Service before the changes take effect. No refund of monthly fees for the remaining current month will be provided in such case.

Article 20 (Severability)

If any provision of these Terms is found to be invalid or unenforceable under applicable law, the remaining provisions shall continue in full force and effect. The invalid or unenforceable provision shall be reinterpreted to the extent permitted by applicable law to reflect as closely as possible the original intent of the parties.

Article 21 (Governing Law and Jurisdiction)

  1. These Terms shall be governed by and construed in accordance with the laws of Japan.
  2. The Tokyo District Court shall have exclusive jurisdiction as the court of first instance for all disputes arising in connection with these Terms.

Article 22 (Contact)

For inquiries regarding these Terms, please contact us at:

These Terms are effective as of February 5, 2026.