Subprocessors
Last updated: August 28, 2026
FilmSpark.AI uses the third-party providers below to deliver the Service. Some of them process content you upload or generate. Others process account and billing information only.
We update this page before engaging a new subprocessor. Enterprise customers under a Data Processing Addendum receive advance notice of changes and may object, as described in that agreement.
Providers that process your content
| Provider | Purpose | Processing location |
|---|---|---|
| fal (Features & Labels, Inc.) | Primary AI model inference and compute. Most generation models available in FilmSpark are reached through fal. | United States |
| Beeble AI, Inc. | AI-powered relighting and compositing. | United States and Korea |
| World Labs Technologies, Inc. | 3D world and spatial generation. | United States, with service providers operating internationally |
| OpenAI | Story and prompt processing, and AI assistant features. | United States |
| Anthropic | Story and prompt processing, and AI assistant features. | United States |
| Story and prompt processing and AI assistant features, via the Gemini API. | United States | |
| Supabase | Database and file storage. Hosted in the United States (US-East). | United States |
| Amazon Web Services | Application hosting, content delivery, and temporary processing storage. | United States |
Model providers reached through fal. Most generation models available in FilmSpark are fulfilled by fal rather than by a direct connection. The model used for each generation is identified within the product. fal maintains a current list of its own subprocessors at trust.fal.ai/subprocessors.
Onward subprocessors. Beeble and World Labs each engage their own subprocessors, including cloud infrastructure, storage, and content delivery providers. Details are available from each provider.
Processing outside the United States. Beeble operates from the United States and Korea, and World Labs operates internationally through its service providers. Content sent to those models may be processed outside the United States. Enterprise customers may restrict a project to models that process only within the United States; contact us to arrange this.
Providers that process account information only
| Provider | Purpose | Processing location |
|---|---|---|
| Clerk | Authentication and account management. | United States |
| Stripe | Payment processing. | United States |
| PostHog | Product analytics and application error reporting. Receives account identifiers, email addresses, and technical error information. | United States |
| MailerLite | Mailing list management for new account sign-ups, and transactional account email. | United States |
| Sentry | Application error and performance monitoring. Receives technical error information, which may include account identifiers. | United States |
| Google Analytics and Google Tag Manager | Website and product analytics, loaded in your browser. | United States |
| Termly | Cookie consent management, loaded in your browser. | United States |
Browser-loaded providers. Google Analytics, Google Tag Manager and Termly are loaded by your browser when you use the Service rather than called by our servers. Sentry receives error reports from both your browser and our servers. None of these providers receives the content you upload or generate. Web fonts used by the Service are served from our own systems and are not loaded from a third party.
No training on your content
We do not use your inputs or outputs to train, fine-tune, evaluate, or improve any AI model, and neither do the providers we engage to fulfil your generation requests. See Section 4.4 of our Terms of Use.