> ## Documentation Index
> Fetch the complete documentation index at: https://docs.get-hive.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Infrastructure and providers

> Where Hive runs and which service providers it relies on — hosting, database, identity, payments, connectors, AI, search, analytics and monitoring — with regions.

Hive is a cloud service built on established infrastructure providers. This page lists where it runs and which providers are involved, so you can assess it alongside your own vendor policies.

<Note>
  This list describes the product's architecture today and may change as Hive grows. For a formal list of subprocessors or a data processing agreement, [contact us](/help/support).
</Note>

## Where your data lives

Hive's core application and database run in **Frankfurt, Germany**:

| Component | Provider | Region |
| - | - | - |
| API and background workers | Fly.io | Frankfurt (`fra`) |
| Database (Postgres) | Supabase | Frankfurt (`eu-central-1`) |
| Product analytics | PostHog | EU |
| Error monitoring | Sentry | Germany |

The web app is served by Vercel's global edge network; these docs are hosted by Mintlify.

## Service providers

| Provider | Role in Hive |
| - | - |
| **Vercel** | Serves the Hive web app. |
| **Mintlify** | Hosts this documentation site. |
| **Fly.io** | Runs the Hive API and background workers. |
| **Supabase** | Managed Postgres database, with row-level security. |
| **Upstash** | Job queue (QStash) and cache and rate limiting (Redis). |
| **Clerk** | Sign-in, sessions and workspace membership. |
| **Stripe** | Payments, invoices and the billing portal. Hive does not store card numbers. |
| **Nango** | Connector sign-in (OAuth), token storage and connector API calls. |
| **OpenAI, Anthropic, Google, xAI, Microsoft** | AI models, reached directly or through a model gateway provider. See [AI and models](/security/ai-and-models). |
| **Tavily** | Web search for Ask and for workflow agent steps that use the web search tool. |
| **PostHog** | Masked product analytics (EU). |
| **Sentry** | Error monitoring (Germany). |
| **Slack, Microsoft Teams** | Only if you install the Hive channel, to answer questions in chat. See [Slack and Teams](/integrations/slack-and-teams). |
| **Object storage (S3-compatible)** | Files you upload and files Hive generates. |

The tools you connect — your CRM, accounting system, mailbox and so on — are your own providers. Hive reads from them under the access you grant.

## How it fits together

```mermaid theme={null}
flowchart TB
  U[You, in the browser] --> W[Hive web app<br/>Vercel]
  W --> A[Hive API<br/>Fly.io, Frankfurt]
  A --> D[(Postgres<br/>Supabase, Frankfurt)]
  A --> Q[Jobs and cache<br/>Upstash]
  Q --> K[Workers<br/>Fly.io, Frankfurt]
  K --> N[Connectors<br/>Nango]
  N --> T[Your tools]
  K --> M[AI model providers]
  A --> C[Clerk sign-in]
  A --> S[Stripe billing]
```

See [How Hive works](/developers/how-hive-works) for the full request and job flow.

## Security of the platform

* **Encrypted in transit.** The API only accepts HTTPS.
* **Signed internal traffic.** Jobs between Hive's components and webhooks from its providers are signature-verified.
* **Always-on API.** At least one API machine is always running, and background jobs are retried at least once if they fail.
* **Isolation.** Workspace data is separated by row-level security in the database as well as in application code.

## Frequently asked

<AccordionGroup>
  <Accordion title="Is my data stored in the EU?">
    Hive's database and application servers run in Frankfurt, Germany, and product analytics are processed in the EU. Some providers — for example AI model providers, the connector platform and the tools you connect yourself — may process data in other regions under their own terms. See [AI and models](/security/ai-and-models).
  </Accordion>

  <Accordion title="Does Hive copy all my data out of my tools?">
    No. Hive reads what it needs to detect signals and answer questions, and leaves your systems of record where they are. Connections start read-only, and you choose what each account may do. See [Accounts and autonomy](/integrations/accounts-and-autonomy).
  </Accordion>

  <Accordion title="Who holds the tokens for my connected tools?">
    For managed OAuth connectors, the connector platform (Nango) stores provider tokens and makes calls on Hive's behalf; Hive stores only a reference to the connection. Secrets Hive holds itself — MCP bearer tokens, webhook signing secrets and model keys — are encrypted per workspace.
  </Accordion>

  <Accordion title="Can Hive be deployed in my own cloud?">
    Hive is offered as a hosted service today. If you have specific hosting requirements, [talk to us](/help/support).
  </Accordion>

  <Accordion title="Do you have a list of subprocessors?">
    The table above lists the providers Hive's architecture relies on. For a formal subprocessor list or a data processing agreement, [contact us](/help/support).
  </Accordion>
</AccordionGroup>

## Related

<CardGroup cols={2}>
  <Card title="Security overview" icon="lock" href="/security/overview">
    Controls inside the product.
  </Card>

  <Card title="Data protection" icon="database" href="/security/data-protection">
    Retention and deletion.
  </Card>

  <Card title="AI and models" icon="microchip" href="/security/ai-and-models">
    AI providers in detail.
  </Card>

  <Card title="How Hive works" icon="diagram-project" href="/developers/how-hive-works">
    Architecture for technical readers.
  </Card>
</CardGroup>