AT A GLANCE
- Your playground draft is saved in your browser.
- Requests pass through Cloudflare to Vansa-owned infrastructure.
- Full API inputs and answers are not stored in the routine usage database.
- Images sent to the API are processed in memory and not stored.
- Operational metadata and diagnostic records are stored separately.
- Privacy questions and requests: [email protected].
01 Who we are
Vansa.org ("Vansa", "we" or "us") is operated in Lithuania. This policy covers our website, documentation, playground, account and usage consoles, billing and API. Vansa.org is responsible for the personal data we use to administer and secure these services.
Vansa.org also operates Vansa Bot, an app for Slack offered under the VANSA Labs brand. Its privacy and data flow notice and the Vansa Bot privacy page explain what the app reads, stores and deletes in a Slack workspace. Requests about those records can be sent to [email protected]. This policy covers the Vansa API requests the app sends with a workspace's own API key.
If your organisation sends personal data to the API on behalf of other people, it is responsible for its own privacy notices and lawful use of that data. Contact us to arrange any required data-processing terms before submitting it. This notice does not replace a data-processing agreement.
02 Information we process
- API content
- The context, any images, questions, answer options and model selection you submit, and the results produced in response. This includes text entered in the Vansa Bot website demo and, for workspaces that install Vansa Bot, Slack message text, test text entered by its administrators and the review criteria the app sends with the workspace's own API key. Before sending, the app replaces user and channel mentions and email addresses with placeholders and reduces links to their label or host; names written in a message can remain. We send the request to the model server to calculate its answers.
- Usage & diagnostics
- Request time and identifier, API-key identifier, model, question and token counts, processing time, status, client IP address and error details. Diagnostic messages can include question identifiers or fragments of submitted content.
- API-key records
- Key names, owner contact details where supplied, administrative notes, access limits, issue and expiry dates, revocation status and usage. We also retain credential information needed to authenticate and administer keys.
- Account & sign-in
- Email address, optional profile name, verification status, securely hashed password if set, linked Google/GitHub identity identifiers, sessions, activation/recovery records, security events, sign-up security-check results and the keyed hashes used to decide welcome-bonus eligibility (see 03). Social sign-in requests basic profile and email access; it does not grant access to Gmail messages or GitHub repositories.
- Billing
- Orders, Stripe customer/subscription/payment identifiers, invoices, payment status, monthly and prepaid token balances, promotions and usage allocations. Purchase confirmations retain the applicable terms. Withdrawal requests retain your name, account email, statement, purchase reference, receipt time and resolution. Stripe collects payment and billing details, including address and any tax information you provide. Vansa does not store full card numbers.
- Service email
- Recipient address, message purpose, delivery status and retry records for activation, password recovery, payments, subscription changes and usage notifications. These are service communications, not a marketing subscription.
- Communications
- Your email address and any information or attachments you choose to send when contacting us about access, support or privacy.
- API waitlist
- Your name, email address, developer or business profile, optional company or project name, intended use and signup date. These details are stored on Vansa-owned infrastructure and are available to the Vansa team through the private administration area.
The public API's routine usage database does not store full request bodies or answers. This is separate from browser drafts and diagnostic logs. A request runs model inference; it does not itself train or update the model. Images attached to an API request are processed in memory to answer that request and are not stored: we do not write them to our databases or logs, and the usage record of such a request contains no image data.
Use sample or de-identified content in the public playground. Only submit information you have permission to share, and avoid passwords, private API keys and unnecessary personal details in your input, including in images.
03 Why we process it
We use the information needed to answer API requests, provide and secure accounts, process purchases, manage subscriptions and token balances, send service notices and respond to support enquiries. Where this involves personal data, we rely on performing our agreement with you or taking steps you request before an agreement.
If you join the waitlist, we use your details to review your request and contact you about API access. Joining does not subscribe you to a general marketing newsletter. You can ask to leave the waitlist at [email protected].
We use usage and diagnostic records to manage capacity, enforce access limits, investigate errors and protect the service against abuse. Our basis is our legitimate interest in operating a reliable and secure service, balanced against your rights. Accounting, tax and required financial records are processed to meet legal obligations. We may also retain or disclose information where another legal obligation requires it.
To keep the welcome bonus to one per person, we store keyed hashes, not the plain values, of your normalised email address, its domain, the Google or GitHub account identifier you sign up with, and the network address from which you complete sign-up or activation. We also record the result of the sign-up security check and the bonus decision. A daily cleanup deletes the domain and network hashes from our active database after 30 days. The email and account-identifier hashes are kept with your account record. We use them only to decide bonus eligibility and to investigate abuse, based on our legitimate interest in preventing fraud and misuse. If a bonus is held for review, your account still works. You can ask us to review the decision at [email protected] and object to this processing at [email protected].
If we ask for consent for an optional use, you can withdraw it without affecting processing that already took place. Vansa returns model estimates; how an application acts on those estimates is determined by its developer or operator.
04 Browser storage
The playground uses local storage to remember your draft context, questions and editor settings. These drafts have no automatic expiry. Clearing the site's data in your browser removes them.
The playground starts with shared demo access. If you explicitly choose to use your own API key, it is kept only for the current page and is not saved by the playground in browser storage. Personal keys saved by an earlier version are removed when the playground loads; your draft is kept. On a shared device, clear site data when you finish. The shared demo credential does not create a personal account or a private cloud draft.
With your permission, our public website and documentation use Google Analytics to help us understand visits and improve the site. The Google tag loads only after you choose Allow analytics. It can process page visits, general browser and device information, and cookie identifiers. We disable advertising signals and advertising personalization. Google Analytics is not included in the API, playground, account or usage consoles, or administration pages.
You can decline without affecting access and change your choice using Cookie settings at the bottom of public pages. We remember your choice in local storage for 180 days on each site; analytics cookies are also limited to 180 days. Withdrawing consent stops analytics on that page and removes its analytics cookies. Clearing site data removes your saved choice. Our page configuration omits URL query strings and fragments and sends only the origin of a referring site.
Infrastructure providers may process technical data needed to deliver and protect traffic. A separate session cookie is used for the administration area.
The account and usage consoles use essential session cookies to keep you signed in. The API key entered for console sign-in is not saved in browser storage. We store a hashed session identifier, its expiry and the associated account or API-key identifier. Recovery links and OAuth state are time-limited and validated separately. Sessions expire after 12 hours; signing out invalidates the session.
The console's account-creation views load Cloudflare Turnstile, which may read and store technical information in your browser that is strictly necessary for this security check. We do not enable Turnstile pre-clearance, so the check does not set Cloudflare's clearance cookie. Turnstile is not used for signing in and is not loaded in your signed-in console.
05 Providers & transfers
Vansa is self-hosted: the website, API, model and Vansa Bot run on servers we own and operate in Lithuania, and their databases are stored there. API requests travel from your device through Cloudflare to that infrastructure. Cloudflare provides traffic delivery and protection. We do not send API inputs to a third-party AI inference provider.
When you allow analytics, Google processes analytics data to provide website reports. This optional processing is based on your consent and may involve infrastructure outside your country. See Google's privacy policy and how Google uses information from sites that use its services.
Stripe processes payment and billing information for checkout, subscriptions, invoices and payment security. Google and GitHub provide the basic profile/email information needed when you choose their sign-in option. YeyMail delivers transactional email. These providers do not receive API message content through the sign-in or billing integration. Their roles and processing may also be described in their own privacy notices.
When you create an account, or finish creating one with Google or GitHub, the console uses Cloudflare Turnstile to check that the sign-up is not automated. Turnstile runs code from challenges.cloudflare.com in your browser and processes technical signals about your connection and browser, such as your IP address, TLS fingerprint, user-agent header, the site key and origin of the page, and the results of automated browser checks. To confirm the check, our server sends Cloudflare the resulting check token together with your IP address. The check does not use the contents of the form fields, such as your name or email address; like all traffic to our site, the submitted form passes through Cloudflare's network in transit. Cloudflare processes this data on our behalf to provide the check, and as an independent controller to improve its bot detection; see the Cloudflare Turnstile Privacy Addendum and Cloudflare's privacy policy. We use the result only to protect sign-up and the welcome bonus against automated abuse, based on our legitimate interest in preventing fraud and misuse (GDPR Art. 6(1)(f)). Cloudflare may process this data outside the EEA under the safeguards described in its privacy policy. Turnstile is not used when you sign in to an existing account.
Providers involved in network delivery and communications may process data needed for those functions. Information may also be disclosed where legally required or necessary to investigate misuse or protect people's rights.
Operating from Lithuania does not mean every network or processing step takes place in Lithuania. Provider infrastructure can involve other countries. Contact [email protected] for current processing-location and transfer information before using Vansa for data with specific residency requirements. Where a restricted international transfer is needed, an applicable legal transfer mechanism must be in place.
Links to external services lead to sites governed by their own privacy notices.
06 Retention & security
Browser drafts remain until you replace them or clear local site data. Account, API-key, billing and operational records are stored separately; revoking a key or cancelling a subscription does not automatically erase past request or billing records. Records needed for legal accounting, security or disputes may need to be retained after account closure.
Vansa currently has no fixed automatic deletion schedule for application usage records. Two sign-up records have fixed limits: a daily cleanup deletes the domain and network hashes used for welcome-bonus eligibility from our active database after 30 days (see 03), and if you start creating an account with Google or GitHub but do not complete it, the email address, name and account identifier they provided expire after 15 minutes and are then deleted. Retention must be assessed against the need to administer access, resolve support cases, investigate security incidents and meet legal obligations. Infrastructure logs and backups may follow separate schedules. Contact us to request deletion or details about records connected to your use.
We use access controls, API authentication and HTTPS for the public service. These measures reduce risk, but no system can guarantee complete security. Protect your personal keys and report suspected exposure to [email protected].
07 Your rights
Depending on the applicable law and circumstances, you may request access to your personal data, correction, deletion, restriction or a portable copy, and object to processing based on legitimate interests. Where consent is the basis, you may withdraw it.
Email [email protected] with enough information to identify the records concerned. We may need to verify your identity. Do not send your full API key. For data submitted by another organisation, it may be necessary to direct your request to that organisation.
You can complain to Lithuania's State Data Protection Inspectorate or the competent authority where you live or work. Contacting us first is optional.
The service is intended for adults. If you believe a child has supplied personal data, let us know so we can assess and address it.
08 Changes & contact
We will update this page when our service or data practices change and revise the date above. Where required, we will provide further notice before a material change takes effect.
For questions about this policy or your personal data, contact Vansa.org in Lithuania using the address below.