Roadmap

The roadmap is anchored to a concrete acceptance matrix, not vibes. Every row below is a scenario a harness app must pass on a real device — and every row passes today, validated on both Android (Kotlin) and iOS (Swift).

Acceptance matrix — passed

#	Scenario	Status
1	Native→RN typed query and RN→native typed query (bidirectional round trip).	Passed
2	Stream native→JS (lossless under JS-thread saturation) and JS→native (collection + cancellation reaching the JS producer’s teardown).	Passed
3	State: first-render sync read of a global value; observe on both sides; Unprovided on close; re-provide reset.	Passed
4	Runtime reload mid-stream — provider stream cancelled, no dead-callback pumping, new epoch rewires parked subscriptions, mirrors re-hydrate.	Passed
5	Two simultaneous instances of one feature — instance-scoped providers isolated; host recreation re-keys cleanly.	Passed
6	ServiceLoader discovery in a minified release build (R8 consumer rules proven, Android).	Passed
7	Prod-mode payload with { a: undefined } crosses safely (always-on encoding proven).	Passed
8	Skew simulation — extra member vs older descriptor → single structured warning + METHOD_NOT_FOUND with diff context; feature-detection path.	Passed
9	Kotlin consume from the main thread → fast loud failure, not an ANR.	Passed
10	dump() debug screen shows live bindings / streams / state / epoch — on Android and iOS.	Passed

First adoption — shipped

The first real-world adoption is done, not planned:

1. Adoption & freeze. The legacy action API / ad-hoc appGet* getters / raw native event emitters are frozen — no new actions or modules.
2. First stream. The highest-value, smallest-surface case (a native polling flow) moved to a typed stream.
3. Feature-owned contract. A pilot feature owns a feature-scoped contract, provided natively at Scope.Feature("YourApp.Feature"); shared capabilities like localisation and analytics consume the global example.host; genuinely host-level actions stay on the app’s existing dispatcher. Runtime-proven (a Sync isLoggedIn JS→native returned true; an analytics call fired end-to-end).
4. SPM end-state. The host app iOS integration is pure SPM — no CocoaPods, no C++ interop in the app target (@_implementationOnly import BridgeKit). This is the target end-state for all the host app feature integrations.

The full before/after is in Migrating a feature to BridgeKit.

Platform expansion

Target	Status
Android	Done. Kotlin runtime + codegen, validated on-device.
iOS	Done, at parity with Android. Swift runtime + codegen, runtime-validated with UI parity; the host app end-state is pure SPM.
Web	The open frontier. LoopbackTransport already runs pure-JS providers in-process; a formal web transport is the remaining work.

Where to verify status

The code and the contract are the authority on what’s landed. This page reflects current state; check the package and the changelog before depending on anything described as future.