Farka
Custom trainingModelsExamplesDocsAPISign inCreate a workspace
Sign up
Custom trainingModelsExamplesDocsAPI
Legal · version 2026-09-10

Privacy Policy

Effective September 10, 2026 · SHA-256 3ed57499bbe7970893f90f1a08ad42da5b761c6e8f825632b28b6f817b203242

This Privacy Policy explains how Applifyer, LLC, 131 Continental Dr, Suite 305, Newark, DE 19713, United States, processes personal data in connection with Farka. It is a notice, not a request to consent to processing that is necessary for the service contract or another lawful basis.

  1. Roles and scope

    Customers generally determine the purpose and means of processing uploaded training data, with Applifyer acting on documented instructions. Applifyer acts as controller for account, authentication, security, billing, support, legal, and service-operation data.

  2. Data collected

    We process account and membership details, authentication and security records, datasets and prompts supplied by customers, model configurations and artifacts, usage and billing records, support communications, and limited cookie-free product analytics.

    • Analytics excludes account IDs, emails, typed content, dataset names, prompts, and training material.
    • Session replay, heatmaps, advertising cookies, and cross-site tracking are disabled.
  3. Purposes and legal bases

    We process data to provide and secure the service, execute customer instructions, meter and bill usage, provide support, improve reliability, prevent abuse, and meet legal duties. Depending on context, the legal basis is contract, legitimate interests, legal obligation, or separately requested consent.

  4. Customer model training

    We do not use customer datasets to train a general-purpose model or a model for another customer unless the customer separately requests and authorizes that use in writing.

  5. Processors and recipients

    Configured GPU, storage, payment, identity, email, deployment, analytics, and observability providers receive only data needed for their function. The categories of provider, what each receives, and where each sits are published at farka.ai/subprocessors. Named providers, sub-processor terms, and processing terms for a specific deployment are available through Data control or a signed data-processing agreement.

  6. Data residency

    A training run may be pinned to the European Union or the United States. Pinning is enforced rather than advisory: compute is selected only from capacity resolved to the requested jurisdiction, there is no fallback to capacity outside it, and a run that cannot be placed in region fails instead of relocating. Each run records the jurisdiction it executed in within its training specification and export bundle.

    • Pinned residency is offered only on a deployment whose application, database, logs, and object storage are all declared in that jurisdiction.
    • A deployment that has not declared them offers automatic placement only, which carries no residency commitment and is recorded as carrying none in the quote and in the run evidence.
  7. International transfers

    Location depends on the declared application, database, logging, backup, and object-storage regions of the deployment, and on the residency selected for each training run. Where personal data leaves its region of origin, transfers rely on adequacy decisions or contractual and supplementary safeguards.

  8. Retention

    Account and operational data is kept while needed for the service, security, billing, disputes, or law. Dataset and artifact retention follows workspace controls and quotes. Deleted primary data may remain in bounded backups until scheduled expiry.

  9. Security

    We use encryption in transit, access controls, tenant ownership constraints, secret separation, audit evidence, bounded retention, and incident procedures. No system can guarantee absolute security.

  10. Rights

    Subject to applicable law, individuals may request access, portability, correction, deletion, restriction, objection, withdrawal of consent, and regulator review. Farka does not sell customer training data. Requests can be submitted in Settings, Data control, or by email.

  11. Cookies and analytics

    Essential session cookies authenticate and secure users. Private cookie-free analytics honors Do Not Track and records only bounded page and performance information without query strings or fragments.

  12. Children

    Farka accounts are for people aged 18 or older. Customer datasets may contain information about minors only where the customer has the required rights, lawful basis, and safeguards. This is distinct from allowing a minor to hold an account.

  13. Policy changes

    We notify users of material changes. A material version is shown in-product for acknowledgement before protected product use continues; acknowledging the notice does not convert contract-based processing into consent.

  14. Contact

    Privacy questions and rights requests may be sent to edihasaj@gmail.com.

Farka

Forge sharp, specialized models from your own data.

ProductCustom trainingModelsExamplesChangelog
DevelopersDocumentationAPIDocs Markdown
TrustSecurityData controlProcessorsPrivacy
LegalTermsAcceptable useRefunds
© 2026 Farka · operated by Applifyer