Engine

Jennefer Engine is a background service without a window. Every Jennefer app talks to it on your computer. In Task Manager it shows as Jennefer Engine.

What it does#

• Runs your local models. The local runtime and your models are shared by every app. See Models.
• Answers every model request. Each app sends its model requests to the Engine with the model and effort its window picked. Without a pick, the Engine uses the app default, then the Hub default. It passes the request to a local model, or to your approved endpoint with the key added. See Model and effort per app.
• Counts tokens of every request for Hub > Usage, with the estimated cost of endpoint requests.
• Shares the GPU. One app at a time gets the GPU, in a queue. See GPU.
• Makes images and videos for Studio with your packs.
• Runs MCP servers for the IDE. See MCP servers.
• Keeps your agent teams that you edit in Hub > Agents. See Agent teams.
• Indexes your library and passes files between apps with Send to. See Library.
• Holds the sign-in for every app on this computer.
• Serves connected apps on a local OpenAI-compatible address, when you turn it on. See Connected apps.
• Links your phone to the model on this computer, when you turn it on. See Mobile.

When it runs#

• The Engine does not start with Windows. The first app or Hub window that needs it starts it.
• It stays awake while an app window is open or a job runs: a download, a pack install, a generation, a chat request or an MCP tool call.
• About two minutes after the last app closes and the last job ends, it goes to sleep. It stops the local model, the image engine and the MCP servers when it does.
• It stays awake while the Hub runs and Connected apps has an allowed app, phone access is on, or a Hangar bot is launched. Quit the Hub from the tray to let it sleep.
• Stop in Hub > Settings > Jennefer Engine stops it and frees the GPU and memory. It stays off until you choose Start or open a Jennefer app.

The Hub shows the Engine's state at the bottom of the sidebar: Engine awake, Engine working, Waking the Engine, Engine asleep or Engine unreachable. Hub > Settings > Jennefer Engine shows its version, the open app windows, the jobs running and the local runtime.

1. 1

   Restart the Engine

   In Hub > Settings, choose Restart. This stops every model and download, then starts the Engine again.

Ports#

The Engine listens on your own computer only (127.0.0.1). Other computers on your network cannot reach it.

If another program uses port 48486, the Engine cannot start. See Troubleshooting.

Files and logs#

The Engine keeps settings and small files in the .jennefer folder of your user profile. Models, Studio packs and the local runtime follow the Jennefer folder you picked in Setup or in Hub > Settings > Storage: <folder>\Data\models, <folder>\Data\packs and <folder>\Data\runtime. Without a picked folder they stay in .jennefer.

The logs are engine.log, ollama.log for the local model and sd-server.log for the image engine. Open the folder from Hub > Settings > Logs with Open folder. Your sign-in and your endpoint key are kept in Windows Credential Manager, not in these files. A key leaves this computer only when you turn on Use from your phone without this computer for a model.

FAQ#