1. How to read this list
Skyller is provided by SKILLS IT SOLUCOES EM TECNOLOGIA LTDA, trading as Skills IT, CNPJ 05.699.856/0001-10. This page presents services that may participate in data processing, grouped by purpose. The recipients for each operation depend on the features the customer uses.
Use depends on the feature, configuration, model and route, including fallback. For a specific assessment, ask support about the recipients and conditions for your intended processing before enabling it.
2. Services that may process content on a customer's behalf
| Integrated service | Purpose | Potential data and conditions of use |
|---|---|---|
| Anthropic | AI inference | Prompts, context, supported files and outputs, according to model and route |
| OpenAI | AI inference | Prompts, context, supported files and outputs, according to model and route |
| AI inference | Content submitted to the model, according to model and route | |
| xAI | AI inference | Prompts, context and outputs, according to model and route |
| DeepInfra | Hosting and inference | Prompts, context and outputs, according to model and route |
| OpenRouter | Model routing | Prompts, context, parameters and outputs, according to route |
| DeepSeek | AI inference | Prompts, context and outputs, according to model and route |
| Z.AI | AI inference | Prompts, context and outputs, according to model and route |
| Alibaba Cloud / DashScope | AI inference | Prompts, context and outputs, according to model and route |
| Xiaomi | AI inference | Prompts, context and outputs, according to model and route |
| Firecrawl | Web reading and extraction | Queries, URLs and returned content when the tool is used |
| Exa | Semantic web search | Queries, URLs and results when the tool is used |
| Serper | Web search | Queries and results when the tool is used |
Queries and URLs may contain personal data provided by users; publicly accessible content is not necessarily free of rights or restrictions. Additional tools connected by customers depend on their permissions and contracts.
3. Identity, billing and website services
The following services support account management and website operation. Each provider's responsibilities depend on its processing role and applicable contract.
| Service | Purpose | Potential data and conditions of use |
|---|---|---|
| Social login | Identity, email and authorized attributes when selected | |
| Microsoft | Social or business login | Identity and authorized organizational attributes when selected |
| Asaas | Billing and payment | Tax, billing and transaction data when used |
| Cloudflare Turnstile | Abuse protection | Technical browser and challenge data when configured and used |
| Vercel | Marketing website hosting and delivery | Web requests, IP addresses and technical access data when visiting the website |
| Vercel Web Analytics | Website access metrics | Technical browsing data on pages with the integration |
| Vercel Speed Insights | Website performance | Technical loading and browsing metrics on pages with the integration |
| Google Analytics | Browsing and usage analytics | Browsing data on pages and interfaces where configured, enabled and authorized |
Payment providers used depend on billing country, currency and methods available at purchase. Newly used providers will be identified before the relevant processing; availability is not determined by website language.
Preferences under the Cookie Policy apply across the website, signup and platform. Vercel Web Analytics and Speed Insights operate on Vercel-hosted pages with permission for their respective category. Google Analytics requires configuration and permission. Optional first-party platform usage and performance metrics follow the same categories; records necessary for operation, security and billing have their own purposes.
4. Internal components
PostgreSQL, Redis, Weaviate, Keycloak, Langfuse, LiteLLM, OCR/Docling and local models form part of the architecture operated by Skyller. These software components are presented separately from external providers receiving data.
5. Countries, retention and safeguards
Skyller and the model API providers it selects and contracts do not use customer inputs, including prompts, messages and documents, or generated outputs to train artificial intelligence models. This condition applies to model routes offered by Skyller, including fallback. Location and operational retention vary according to the service, account, endpoint and contract used. Independent tools and services connected and contracted directly by the customer are governed by their own conditions.
The DPA sets out subprocessor selection, contracting and oversight obligations. The Privacy Policy describes processing and the channel for information about recipients, retention and transfer mechanisms.
6. Changes and contact
For DPA-covered processing, Skyller will communicate additions or replacements of subprocessors before use, allowing substantiated objections under the Agreement. Processing will remain limited to authorized purposes and the scope of features used.
Contact: contato@skillsit.com.br. The three language versions describe the same scope of services.