Acceptable Use Policy (AUP)

Version 1.0.0 of {{EFFECTIVE_DATE}}.

1. No personal data in questions and facts#

1.1. Do not enter personal data of individuals (names, phones, emails, addresses, ID data, health data, etc.) into question texts, brand names, facts, tags or other fields. Buyer questions and brand facts must not identify a person. The exception is business contact data of Users in accounts.

1.2. Question and fact texts are sent to third-party AI providers, including outside Russia (via an LLM aggregator), so personal data in them is not allowed. A violation makes the Customer liable as operator.

1.3. Enter a brand name that coincides with a sole proprietor's personal name only if it is publicly used as a brand; where possible use a de-identified name.

1.4. Do not enter special-category personal data (health, beliefs, etc.) or biometric data.

2. Prohibited content#

Do not use the Service for questions, facts or purposes that: violate law (including on extremism, terrorism, pornography, drugs, unlawful gambling, child protection); infringe third-party rights (IP, secrecy, reputation); constitute unfair competition, defamation or misleading advertising; contain malware; or seek answers about illegal goods and services.

3. Abuse of AI providers#

3.1. Do not use the Service as a free or bypass channel to AI models: general chat, content generation, API proxying, resale of tokens.

3.2. Do not try to bypass provider or plan limits, technical restrictions or safety filters; do not make requests violating provider terms (Yandex, Sber, model providers via the aggregator), including the GigaChat acceptable-use policy, which also applies to you (no resale of access, no circumvention, no reputational harm, etc.).

3.3. Do not use AI answers obtained through the Service to train or improve models competing with the providers or the Service.

3.4. Do not inject instructions into questions or facts to manipulate the judge model or answers (prompt injection).

4. Protecting the Service#

Prohibited: automated extraction from the Service (scraping) outside the official API and your plan; bypassing authentication or access limits; reverse engineering, decompiling or seeking source code; vulnerability scanning and load testing without written permission; load that endangers other customers; repeated registration to evade the trial or limits; sharing credentials or reselling access (except the Agency plan within its scope).

Responsible disclosure of vulnerabilities: security@kleogeo.ru.

5. Limits#

Plan limits (questions, providers, repeats, projects, tokens) and API and sign-in rate limits apply. Beyond them operations are rejected (codes plan_limit_exceeded, too_many_attempts). We may temporarily throttle to protect stability.

6. Publishing results#

The Customer is responsible for publishing reports and statements about third parties and competitors, must keep "AI-generated" labels and disclaimers in reports it passes on, and must not present AI answers as statements of the Provider or the AI provider.

7. Sanctions and country restrictions#

Do not use the Service for sanctioned persons, from countries in clause 18.2 of the offer, or in breach of export control; Users must not access the Service from such jurisdictions.

8. Consequences#

Warning; feature restriction; suspension; termination without refund; removal of violating content; referral to law enforcement where required. Report a violation: support@kleogeo.ru.

Notes for the lawyer#

Check clause 3.2 against the current GigaChat acceptable-use policy and the aggregator/OpenAI terms. Clause 1.2 reflects the owner's decision that the ChatGPT point runs via an aggregator outside Russia.