> For the complete documentation index, see [llms.txt](https://communityone.gitbook.io/communityone/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://communityone.gitbook.io/communityone/feature-deepdive/self-hosted-deployment.md).

# Self-Hosted Deployment

> Run CommunityOne entirely inside your own infrastructure — your data stays on your network, under your control.

CommunityOne offers **on-premises (self-hosted) deployment** for Enterprise customers (under a custom contract) who need their community data to remain within their own environment.

On-premises deployment is available exclusively under a **Custom** **Enterprise agreement**. [Book a call with our team](https://calendly.com/louisalu) to discuss your requirements.

***

### Why self-host?

For most teams, our managed cloud is the fastest way to get value. But organizations with stricter security, privacy, or governance needs often require that sensitive data never leave their environment. Self-hosting is built for exactly those teams.

* **Data sovereignty** — Your community data, conversations, knowledge bases, and analytics stay within your infrastructure and your chosen region.
* **Network isolation** — Run within your private network, behind your own firewalls and access controls, with no exposure to the public internet beyond the connections you explicitly approve.
* **Meet your compliance obligations** — Keep regulated and sensitive data inside the environment your security and compliance teams already govern, so CommunityOne fits within your existing controls rather than expanding your audit surface.
* **Your AI, your choice** — Connect CommunityOne to your own self-hosted AI models so that prompts, responses, and embeddings never leave your network, or bring your own provider keys when you prefer a commercial model.

***

### Deployment model

Self-hosted CommunityOne is delivered as a set of containers orchestrated with **Docker Compose**, designed to run on a single host that you provide. Because a single-tenant deployment is lightweight, you don't need a cluster or specialized orchestration — a standard Linux host with a container runtime is enough to run the full platform.

You provide the environment; we provide the images, configuration templates, and hands-on guidance to bring everything up and validate it.

* **One entry point.** A single reverse proxy terminates TLS and routes traffic to the dashboards and APIs. All other services stay on an internal network with no exposed ports.
* **You own your data.** Your datastore lives inside your environment. We never receive a copy.
* **Bring your own AI, or bring your own keys.** Point CommunityOne at AI inference and embedding endpoints you host yourself for a fully private setup, or supply your own provider credentials.
* **Configuration-driven.** Endpoints, credentials, and integrations are configured, so the same release can run in either a fully private setup or a bring-your-own-key setup.

***

### Deployed services

A self-hosted deployment runs the same products you already know, packaged as the following services:

| Service                        | Role                                                                                                                                     |
| ------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------- |
| **CommunityOne bot (Discord)** | Connects to your community platform and powers moderation, automation, and the data feed that everything else builds on.                 |
| **AI assistant API**           | Handles AI agent configuration, knowledge, and conversations; assembles context and calls your chosen inference and embedding endpoints. |
| **Knowledge sync workers**     | Keep the AI assistant's knowledge up to date from your connected sources on a schedule.                                                  |
| **Analytics workers**          | Compute community and member analytics on a schedule.                                                                                    |
| **Analytics dashboard**        | The web dashboard where your team explores community and member insights.                                                                |
| **Configuration dashboard**    | The web app where your team configures the AI assistant, knowledge, and platform settings.                                               |
| **Analytics API**              | Read-only API that exposes the same analytics data programmatically.                                                                     |
| **Database**                   | The datastore that ties everything together; runs inside your environment, or connect CommunityOne to a database you manage.             |
| **Supporting services**        | Internal components (task broker, reverse proxy / TLS termination) that support the services above.                                      |

The vector store used by the AI assistant is rebuilt automatically on startup, so it requires no separate backup.

### Network & connectivity

A self-hosted deployment is **network-isolated**: the community bot maintains a live connection to Discord, and that real-time feed is what powers analytics, AI responses, and automation. Outbound connectivity is therefore required and is limited to a small, documented allowlist that your security team can review before deployment:

* **Your community platform** — required, so the bot can receive and act on events in real time.
* **AI inference & embeddings** — internal endpoints you host (fully private), or your chosen provider when bringing your own keys.
* **Container registry** — only at install and upgrade time, to pull images.

Inbound access is limited to the dashboards and APIs, served over TLS using certificates issued by your organization, and reachable only within your network.

### What you provide

* A Linux host (or VM) with a container runtime, sized to your community footprint.
* Outbound access for the allowlist above.
* TLS certificate(s) from your organization for the dashboard and API hostnames.
* Optionally, your own database and/or self-hosted AI endpoints.

We'll right-size the host and confirm requirements with you during discovery.

### Updates & support

Updates are delivered as new container images. Upgrading is a matter of pulling the new release and applying any included data migrations — we provide the steps and support you through them. Enterprise customers receive dedicated support for the life of the agreement.

### Talk to us

On-premises deployment is part of our **Custom** **Enterprise** offering and is set up together with our team — not self-served.

To explore whether self-hosted CommunityOne is right for your organization:

[**Book a demo**](https://calendly.com/louisalu) — or reach out to your CommunityOne contact directly.

We'll walk you through the platform, understand your requirements, and map out what a deployment in your environment would look like.