Last updated: September 16, 2026
Subprocessors
Belong Universe uses service providers to operate Simetrik Agent. This page describes providers used for hosting, AI processing, monitoring and email delivery. Which providers process a particular request depends on the enabled features, selected model and deployment configuration.
| Provider / service | Purpose | Data that may be processed |
|---|---|---|
| Google Cloud | Application hosting, databases, storage, task execution and Google-hosted AI services when selected. | Account identifiers, application records, prompts, conversation and task history, files and operational data. Selected AI services receive relevant model inputs. |
| Anthropic | Claude model processing when selected. | Relevant prompts, conversation context, tool results and document content needed to perform the requested task. |
| OpenAI | Model and voice processing when enabled. | Relevant model inputs and, for voice features, audio and conversation content. |
| Vercel AI Gateway | Routing AI requests to the configured inference provider. | Model inputs and outputs and request metadata. Requests routed through the gateway are also processed by the selected downstream inference provider. |
| Cerebras | AI inference for supported Cerebras model routes when selected. | Relevant prompts, conversation context, tool results and document content supplied to that model route. |
| Grafana Cloud | Operational monitoring and troubleshooting. | Telemetry such as request and task identifiers, timings, errors and diagnostic attributes emitted by the service. |
| Resend | Email delivery for enabled notification and user-requested email features. | Recipient addresses, message contents, attachments where applicable, and delivery metadata for those emails. |
AI routing and connected services
Model processing can be direct or routed through a gateway or cloud AI service. The model and inference provider may differ; a model’s name alone does not identify the company hosting inference. For the provider route applicable to your deployment, contact us using the address below.
Slack and other services that users connect are accessed under the permissions granted for those connections. Authorizing a connector allows Simetrik Agent to exchange relevant information with that service to perform the requested task. Those connected services also have their own privacy terms.
The Slack profile email permission is used for account identification and user-requested Slack actions. It is not used for email marketing. Email delivery through Resend is a separate feature.
Retention and location
Application conversation and task history is separate from records held by a service provider. Retention and processing location depend on the applicable service, deployment and account configuration. This page does not promise zero retention or a single processing country across all providers. Contact us for deployment-specific requirements before supplying data that needs a particular location or retention arrangement.
Changes and questions
This page records the providers and processing purposes described above as of its update date. Questions about subprocessors, a selected model route or deletion of retained Simetrik Agent data can be sent to davi.suga@simetrik.com. Include your organization and account email, but do not include passwords, access tokens or sensitive task contents.