---
title: "Deployment: managed service and self-hosted options"
description: "Run the Duale AI platform as a managed service in the EU or self-hosted in your own infrastructure with the same SDK."
lang: en
lastUpdated: 2026-10-23
url: https://duale.ai/en/product/deployment
---

<Hero
  eyebrow="Deployment"
  title={
<>
  Run your AI agent platform <AccentText>in the EU, or in your own infrastructure on request</AccentText>
</>
}
  subtitle="Choose the managed service, operated by Duale AI in the European Union, or a self-hosted Platform in your own infrastructure. Either way, your teams delegate the same work and get back one result they can use."
  primaryCta={{ label: "Plan your first use case with us", dialog: "contact" }}
  secondaryCta={{ label: "Open the Trust Center", href: "https://duale.ai/en/product/security.md" }}
/>

<Section title="Choose who runs the Platform" subtitle="Let Duale AI operate it for you, or run it where your data already lives.">
  <CardGrid columns={2}>
    <Card title="The managed service" icon="globe">
      Start with no infrastructure to set up. Duale AI runs the Platform, hosts your data in the European Union and
      charges a fixed price per Task.
    </Card>

    <Card title="Your own infrastructure" icon="layers">
      Run the same Platform today in your own cloud or data center. You and Duale AI agree who operates what.
    </Card>
  </CardGrid>
</Section>

<Section title="Integrate once. Choose where the work runs." subtitle="Your applications send the same Task to the managed service or to the Platform in your infrastructure.">
  ```mermaid
  flowchart LR
      accTitle: The same delegated work, wherever the Platform runs
      accDescr {
        Your applications send the same Task to the managed service in the
        European Union or to the Platform in your own infrastructure. Both use
        the model providers you choose, and both return one result your team
        can use.
      }

      Apps["Your applications"]
      Managed["Managed service in the European Union"]
      Yours["Platform in your own infrastructure"]
      Providers["Your model providers"]
      Result["One result your team can use"]

      Apps -->|the same Task| Managed
      Apps -->|the same Task| Yours
      Managed --> Providers
      Yours --> Providers
      Managed --> Result
      Yours --> Result
  ```

  **Illustrative example.** An industrial group starts market and competitor research on the managed service, with
  nothing to install. Its legal team wants confidential contracts to stay in the group's own data center. The group runs
  the Platform there for contract review. Both teams build on the same [Python SDK](https://duale.ai/en/docs/sdk.md) integration. Each team
  gets back:

  - market and competitor research, on the managed service: a dated brief that cites every source;
  - contract review, in the group's data center: each contract with the clauses that depart from the legal playbook.
</Section>

<Section title="The same models and agents in both places" subtitle="Wherever the Platform runs, your teams keep the providers and agents they rely on.">
  <CardGrid columns={2}>
    <Card title="Your model providers" icon="mixer-horizontal">
      Connect the providers you choose. Each Task runs on the models you enabled.
    </Card>

    <Card title="Each agent with its own identity" icon="person" href="https://duale.ai/en/product/identity-and-access.md">
      Add as many agents as you need. Each one gets only the access you grant.
    </Card>
  </CardGrid>
</Section>

<Section title="Bring the Platform into your infrastructure" subtitle="Your deployment starts from your first use case and an installation agreement built around your environment.">
  <CardGrid columns={3}>
    <Card title="Share your first use case" icon="chat-bubble">
      Tell us which work your teams want to delegate first and where its data must stay.
    </Card>

    <Card title="Agree the installation" icon="file-text" href="https://duale.ai/en/docs/security/shared-responsibility.md#run-the-platform-in-your-infrastructure">
      Your team and Duale AI set the scope, the split of operations, support and commercial terms in one installation
      agreement.
    </Card>

    <Card title="Get your first results" icon="rocket">
      Your teams send their first Tasks through the same SDK as on the managed service.
    </Card>
  </CardGrid>
</Section>

<Section title="Place each use case where its data belongs" subtitle="Keep sensitive work in your own infrastructure and start the rest on the managed service.">
  <CardGrid columns={2}>
    <Card title="Legal" icon="reader" href="https://duale.ai/en/use-cases/legal.md">
      Review each contract against your own playbook while your contracts and evidence stay in your infrastructure. For
      general counsel, legal operations and compliance.
    </Card>

    <Card title="R&D" icon="lightning-bolt" href="https://duale.ai/en/use-cases/research-and-development.md">
      Run literature syntheses and case pre-reads while your unpublished research and case files stay in your
      infrastructure. For research directors and scientific information teams.
    </Card>
  </CardGrid>
</Section>

<Section title="Before your first deployment">
  <CardGrid columns={3}>
    <Card title="One fixed price per Task" icon="check">
      A standard Task costs €0.10 and a fast Task €0.20, excluding VAT. [See pricing](https://duale.ai/en/product/pricing.md)
    </Card>

    <Card title="Your providers, your contracts" icon="layers">
      Your model providers bill you directly for inference.
    </Card>

    <Card title="Encrypted storage" icon="lock-closed">
      All stored data is encrypted, from Library files to database backups.
    </Card>

    <Card title="Your content remains yours" icon="file-text">
      Duale AI does not reuse your content to train a model. [See subprocessors](https://duale.ai/en/legal/subprocessors.md)
    </Card>
  </CardGrid>
</Section>

## Deployment questions

### Can we start on the managed service and bring a use case in-house later

Yes. Your teams keep working on the managed service until the use case runs in your infrastructure. [See
customer-hosted terms in Pricing](https://duale.ai/en/product/pricing.md), then [see what the move means for your applications in
Architecture patterns](https://duale.ai/en/docs/model-routing/architecture-patterns.md#separate-cloud-and-customer-hosted-work).

### How much notice do we get before planned maintenance on the managed service

48 hours ahead, by email. [See availability in the terms of sale](https://duale.ai/en/legal/cgv.md#availability).

### Will we know before a new subprocessor handles our data on the managed service

Yes. Duale AI emails you at least 30 days before it adds or replaces a subprocessor, and you can object. [See the
current subprocessors and the notice rules](https://duale.ai/en/legal/subprocessors.md#changes).

### Can each subsidiary keep its own providers, costs and history, on the managed service or in our infrastructure

Yes, and each subsidiary's access stays apart from the rest of the group too. [See the figures each subsidiary reads
in Usage analytics](https://duale.ai/en/product/usage-analytics.md).

<Cta title="Get your first use case running, managed or in your infrastructure." primaryCta={{ label: "Plan your first use case with us", dialog: "contact" }} secondaryCta={{ label: "Open the Trust Center", href: "https://duale.ai/en/product/security.md" }} />

## Related content

- [Security controls for your agent platform](https://duale.ai/en/product/security.md)
- [AI that takes on real work, with and for your people](https://duale.ai/index.md)
- [Shared responsibility: platform and customer duties](https://duale.ai/en/docs/security/shared-responsibility.md)
- [Product: Task execution and platform capabilities](https://duale.ai/en/product.md)
- [Terms of Use: platform rules and responsibilities](https://duale.ai/en/legal/cgu.md)
- [Customer service: calls reviewed, trends traced to accounts](https://duale.ai/en/use-cases/customer-service.md)

---

## Sitemap

See the full [Markdown sitemap](https://duale.ai/sitemap.md) for all pages.