Tauri state between crates

[alexiokay]

I want to separate my tauri app into different crates like: core, to allow user to run the app as server headless, but the core is deeply integrated with tauri app_handle persisted state (shared_state.rs)

The core can run itself without tauri state, but if I want to still integrate it with frontend I must somehow persist some things. I don't want to use memory for that not Files or SQLite.

The question is: how? I'm quite new to rust.

[FabianLars]

I must somehow persist some things. I don't want to use memory for that not Files or SQLite.

tauri::State is not persistent, it's also just stored in memory.

I typically keep core crates completely uncoupled from tauri crates so i can use them in non-tauri apps without pulling in so many dependencies.

I don't think i can really offer any suggestions especially without further info i guess, just wanted to say what i said above.

[alexiokay]

Sorry I meant state stored in memory.

Overall best practices for writing apps with decoupled/independed crates from tauri?

So my main crate has folders/modules like:

• mcp client
• system controls (windows api, etc)
• flow control (visual flow builder api - runner etc.)
• macros creator and runner
• websocket server which gets messages and run other modules or their combinations.
• http server
• auth
• shated state
• some common utils
• mdns server

So the core is supposed to run:

• macros,
• some system functions,
• mcp tools from mcp servers
• external apps integrations

But when the core will be deployed on cloud, then won't be able to run system tools directly.

So i really wonder what my target architecture should look like and how am I suppose to decouple the core crate from tauri to run the flows headless and have seamless Integration into the tauri app.