Stream: git-wasmtime

Topic: wasmtime / issue #13705 C API support status for Async Co...


view this post on Zulip Wasmtime GitHub notifications bot (Jun 22 2026 at 07:20):

rudikone opened issue #13705:

Hello.

I'm evaluating Wasmtime as an embedding runtime from Java through the public Wasmtime C API.

I noticed that Async Component Model support (component-model-async) appears to be available in the Rust API, including APIs such as:

At the same time, I cannot find equivalent functionality in the public C API.

From the documentation it looks like the C API currently exposes the older async mechanism based on:

which solves host-side asynchronous execution, but does not appear to expose the Async Component Model scheduler or guest task APIs.

Questions:

  1. Is exposing Async Component Model functionality in the C API part of the current roadmap?
  2. Is there already an issue tracking support for:

    • call_concurrent
    • run_concurrent
    • concurrent guest Tasks
    • component-model future / stream
    • func_wrap_concurrent
  3. Is the current recommendation for non-Rust embedders to:

    • use Rust as an adapter layer and expose custom FFI,
    • or wait for first-class C API support?
  4. Are there known architectural blockers preventing a direct mapping of the Async Component Model APIs into the C API?

The motivation is support for WASI 0.3 style component composition from Java without introducing an intermediate Rust host layer.

Thanks.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 29 2026 at 14:00):

alexcrichton added the wasmtime:c-api label to Issue #13705.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 29 2026 at 14:00):

alexcrichton added the wasm-proposal:component-model-async label to Issue #13705.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 29 2026 at 14:01):

alexcrichton commented on issue #13705:

Yes *_concurrent functions aren't exposed just yet in the C API. There'll need to be some work necessary to get those exposed, but my hope is that it's relatively similar to StoreContextMut<'_, ..> and how it's exposed. No one's lined up to do this just yet, however.


Last updated: Jul 29 2026 at 05:03 UTC