Privacy Policy

This Policy applies to the Vetta Services we operate, including the website, sign-in and account center, downloads and updates, subscriptions and billing, client features that communicate with our servers, and optional cloud sync, usage metering, and model routing you enable.

This Policy does not apply to third-party model providers you configure yourself, plugins, MCP servers, third-party marketplace content, external tools or websites, or third-party-maintained open-source components. Those parties have their own privacy terms; we do not control their processing.

To request our full legal entity name, registered address, or a copy of this Policy, email [email protected]. “Personal data” means information that identifies or can reasonably be linked to an individual.

• “Input”: prompts, instructions, files, code, knowledge-base content, tool results, and similar content you submit.
• “Output”: text, code, summaries, suggestions, structured results, and other content returned by the Services.
• “User Content”: Input, Output, and related conversations, projects, and configuration stored locally or (if enabled) in the cloud.
• “Account Data”: email, authentication identifiers, subscription and billing metadata, and similar account records.
• “Local Data”: conversations, file indexes, knowledge bases, API keys, and preferences stored on your device by default and not uploaded to our servers merely by installing the app.
• “Service Providers”: processors that handle data on our instructions (e.g., hosting, payments, email, authentication).

We collect personal data only as needed, lawfully and for stated purposes. Depending on how you use the Services, this may include:

• Account and identity: email address; email verification codes / password hashes (if you set a password); identifiers, display name, avatar, and email from GitHub, Google, or other sign-in providers within the scope you authorize.
• Subscription and transaction data: plan type, order IDs, billing cycle, payment status, and invoice-related fields. Card numbers and similar sensitive payment details are handled by payment processors; we do not store full card numbers.
• Usage and billing metadata: model call counts, token/credit usage, timestamps, selected model or route IDs, HTTP status and error codes—for billing, quotas, and abuse prevention (typically without full conversation text).
• User Content (only when a feature you use requires upload or forwarding): if you use models we host or proxy, Input may be sent to the model provider to generate Output; if you enable cloud sync, feedback, or diagnostics, related content may be uploaded per your choice.
• Device and log data: IP address, device type, OS and client version, browser type, language, crash/performance logs, and security/access logs.
• Cookies and similar technologies: session continuity, preferences, security, and limited analytics (see below).
• Support communications: content and attachments you send by email or support channels.
• Marketing preferences (if you subscribe): email address and unsubscribe status.

We do not intentionally collect government ID numbers, biometrics, or precise real-time location as a product requirement. Please do not submit other people’s sensitive personal data or data you are not allowed to process.

The Vetta desktop client is local-first by default: chat history, project files, knowledge bases, your own model API keys, and plugin/skill configuration are stored on your device and are not automatically uploaded to our servers merely because you install and use the app offline.

• Third-party model API keys you enter: stored locally in the client; we do not collect or require you to send us those keys.
• Conversations and knowledge-base content that remain local and for which you have not enabled cloud sync or diagnostic upload.
• Sensitive device permissions (contacts, photos, etc.) beyond what a feature needs and what the OS has authorized.

When you connect your own models, self-host, or run inside a private network, data flows follow your configuration; we cannot control third parties on those paths. See “Third-party models and integrations.”

We process personal data for the following purposes:

• Provide and maintain the Services: registration and sign-in, workspace features, downloads/updates, subscription entitlements, and quotas.
• Generate AI responses: forward necessary Input to the model you select or that we integrate, to return Output.
• Security and abuse prevention: detect fraud, account takeover, attacks, and policy-violating use.
• Billing and support: process payments, invoices, disputes, and support requests.
• Product improvement (not training on your User Content): use aggregated or de-identified usage, crash, and performance data to improve reliability and UX.
• Communications: transactional notices (security, billing, material policy changes); marketing only with consent or as otherwise allowed, with opt-out.
• Legal compliance: respond to lawful requests and mandatory obligations.

We do not use your User Content (Input and Output) to train or improve general-purpose foundation models. If that ever changes for an optional feature, we will notify you prominently and obtain consent or provide a clear opt-out, as required.

To provide AI features, Input may be sent to a model provider you choose or to model routing/APIs we operate. Responsibilities differ by path:

• Bring-your-own key / direct third-party: your device or gateway talks to that provider; their privacy policy and terms govern, and risk sits in your relationship with them.
• Hosted or proxied model calls: we and our infrastructure providers may process request metadata and Input needed to produce Output—only to provide the Service, security, billing, and troubleshooting.
• Plugins, MCP, skills, and external tools: enabling them may send content to third parties per each extension’s design; review permissions before enabling.
• Third-party sign-in (GitHub, Google, etc.): we receive only authorized profile fields and follow platform rules (including the Google API Services User Data Policy’s Limited Use requirements). We do not use Google user data for ads or sell it.

Model providers may include OpenAI, Anthropic, Google, Alibaba Cloud (Qwen), DeepSeek, and others depending on your selection and our current integrations. Check each provider’s site for the latest policy. We aim to document integrations in-product or in docs; the list may change over time.

We do not sell your personal data and do not sell your conversations to advertisers. We share data only in these cases:

• Service Providers: cloud hosting and storage, CDN, payments, email delivery, authentication, error monitoring—under contracts and only as needed to provide the Services.
• Model and AI infrastructure providers: as described above, to generate Output or maintain routing.
• Legal and compliance: when required by law, valid legal process, or to protect the rights, property, and safety of us, users, or the public.
• Business transfers: merger, acquisition, financing, or asset sale; the successor will remain bound by this Policy or equivalent protections, with notice where appropriate.
• With your direction or consent: e.g., export, share links, or support escalations you authorize.
• Affiliates: under common control, only for purposes in this Policy and with equivalent safeguards.

On the website and web surfaces we use cookies, local storage, and similar technologies:

• Essential: sign-in sessions, security (e.g., CSRF), load balancing—core features may not work without them.
• Functional: language, theme, and similar preferences.
• Analytics (if enabled): aggregated understanding of traffic and product use.

You can refuse or delete cookies in your browser; blocking essential cookies may prevent sign-in. We do not use cookies for cross-site advertising profiles.

We retain personal data only as long as needed for the purposes in this Policy, then delete or anonymize it, unless law requires longer retention.

• Account Data: kept while your account is active; after deletion, removed or anonymized within a reasonable period (typically within 30 days), except where law requires longer retention (e.g., tax, disputes).
• Conversations and User Content: Local Data is managed on your device; if you use our cloud storage/sync, after account or item deletion production copies are typically purged within 30 days, with backups rolling off on a slightly longer cycle.
• Security and access logs: generally shorter periods (weeks to months) for security and troubleshooting.
• Billing and transaction records: retained as required by tax and accounting rules.
• Ephemeral processing for model calls: we do not keep full prompts as long-term training corpora; temporary caches/queues last only as needed for delivery, retries, and safety checks.

We may process and store information in the regions where you use the Services and where we or our providers operate. For cross-border transfers we use reasonable safeguards (contractual terms, access controls, encryption in transit) and, where required, additional legal mechanisms.

If you send data to an overseas model provider or endpoint you configure, that transfer is initiated by you. For our current primary hosting regions, email [email protected].

We implement technical and organizational measures aligned with industry practice, including:

• TLS in transit and appropriate protection at rest where applicable;
• role-based access control, least privilege, and operational logging;
• credential management and environment isolation;
• dependency and infrastructure updates and vulnerability response;
• backups and continuity measures where feasible.

No method of transmission or storage is perfectly secure. Protect your passwords, API keys, and devices; enable OS updates and disk encryption. If you suspect a breach, contact [email protected] immediately.

Depending on your location (e.g., China’s PIPL, EU/UK GDPR, California CCPA/CPRA), you may have rights to access, copy, correct, delete, port, restrict or object to processing, withdraw consent, and lodge a complaint. You can exercise rights by:

• Account settings: view/update profile basics and manage billing after sign-in.
• Desktop client: delete chats, projects, knowledge bases, or uninstall to clear Local Data (feature-dependent).
• Account deletion: via settings or email request; some legally retained records may remain.
• Marketing opt-out: unsubscribe links or reply to opt out.
• Requests: email [email protected] with the request type and verifiable account details; we respond within applicable deadlines (often 15–30 business days depending on law).

We may verify identity before acting. We may decline requests that are manifestly unfounded, excessive, or not required by law, with an explanation. Because AI Output is probabilistic, we cannot guarantee “correction” of every factual statement in Output into a single ground-truth answer.

The Services use machine-learning models to generate Output from your Input. Output may be inaccurate, incomplete, or biased, and is not an automated decision that alone produces legal or similarly significant effects (e.g., credit, hiring, or medical diagnosis). Do not rely solely on Output for high-impact decisions; consult qualified professionals where appropriate.

We may use automated systems to detect abuse, fraud, or security threats and to limit access. If you believe a restriction was in error, contact support to appeal.

The Services are for adults. You must be at least 18 years old to create an account. We do not knowingly collect personal data from anyone under 18. If you are a parent or guardian and believe a minor provided data, contact us and we will delete it and restrict the account.

EEA/UK/Switzerland: legal bases may include contract performance, legitimate interests (e.g., security and product improvement where they do not override your rights), consent (e.g., marketing), or legal obligation. You may complain to your local supervisory authority.

China: we process personal information under the PIPL and related rules. Where separate consent is required (e.g., certain sensitive data, disclosure, or cross-border transfer), we will obtain it through product flows or writing unless an exception applies.

California and similar US state laws: you may have rights to know, delete, correct, port, and limit “sale” or “sharing” for cross-context behavioral advertising. We do not sell personal information. Submit requests via the contact email; we will not discriminate against you for exercising privacy rights.

We may update this Policy to reflect product, legal, or practice changes. The updated Policy will be posted here with a new effective date. If changes materially affect your rights or our processing, we will provide reasonable advance notice via in-product notice, client prompt, or email where appropriate. Continued use after the effective date constitutes acceptance, except where law requires fresh consent.

For questions, complaints, privacy rights requests, or product support: [email protected]. We will respond within a reasonable time.