How hosted Genfeed Cloud and the self-hosted Community edition differ in operations, tenancy, billing, generation and support.
Read the guide, then turn it into a concrete plan for your brand.
Genfeed ships as a hosted product, Genfeed Cloud, and as a self-hosted Community edition. They run the same codebase, which is open source under AGPL-3.0, but they differ in who operates it and what is available out of the box.
| Topic | Genfeed Cloud | Community self-hosted |
|---|---|---|
| Who runs it | Genfeed operates the service at app.genfeed.ai | You run Docker, storage, backups and updates |
| Organizations | Multiple isolated organizations, with an organization switcher | One organization per deployment |
| Brands | Many per organization | Many |
| Sign-in | Email and password, magic link, or Google | No login by default; an optional login wall adds local accounts |
| Generation | Managed models paid with credits; your own provider keys on paid plans | Your own provider keys; no Genfeed credits |
| Credits and plans | Organization billing and subscriptions | No billing locally; managed credits can be bought on Cloud and used through the API |
| Storage | Managed by Genfeed | Local filesystem and your own PostgreSQL |
| Team governance | Members, roles and invitations per organization | Single-tenant account controls |
| API and MCP | REST and MCP at
api.genfeed.ai and mcp.genfeed.ai | REST and MCP on your own host |
| Support | Contact Genfeed; terms depend on your plan | Community channels on GitHub |
Start at app.genfeed.ai/sign-up or follow Getting started with Genfeed Cloud. See pricing for plans.
Install instructions are in the self-host documentation, and genfeed.ai/self-hosted summarizes what you get. The fastest start is one command:
npx @genfeedai/create my-genfeedSelf-hosting costs your infrastructure plus whatever your model providers charge. Cloud prices, included credits and support terms are on genfeed.ai/pricing. Neither edition promises a support response time in the repository.
A self-hosted instance does not automatically sync to Cloud. A Cloud API key authorizes only the calls you make to Cloud, such as managed inference. There is no assumed complete migration path for workflows, brands, templates or assets in either direction, so check export and import support for each resource before planning a move.
When a self-hosted install uses an external model provider, prompts and media leave your deployment under that provider's terms. Keeping everything local requires local model infrastructure as well.