Rust Implementation of Apache Burr

[jecsand838]

Is your feature request related to a problem? Please describe.

I’d like to use Burr’s state-machine / action-graph approach in environments where Rust is the primary deployment/runtime (e.g., Rust microservices, single-binary distribution, memory-safety requirements, low-latency workloads, WASM targets, etc.).

Currently this means either running Burr as a separate Python service, embedding Python, or re-implementing similar functionality ad-hoc, each of which adds operational complexity and makes it harder to standardize on Burr as a core orchestration primitive across a Rust-heavy stack.

Describe the solution you'd like

I’d like to propose (and happily help implement) an officially supported or officially endorsed Rust implementation of Burr with the goal of full functional parity with the existing Python implementation, including:

• Core runtime parity: actions, transitions, state semantics, sync/async execution, parallelism features, hooks, typing system behavior, etc.
• Persistence & tracking parity: same data model and on-disk formats where applicable, so recorded runs are comparable/portable.
• Telemetry/observability parity: compatible tracing/metrics behavior (and the same conceptual execution events).
• Tracking server + UI parity: ideally, the existing UI can be used unchanged by matching the server API contract.
• CLI parity: equivalent commands/subcommands and behavior, or a clearly documented mapping.

Proposed Implementation approach:

• Define the spec as golden/characterization tests derived from the reference implementation (recorded runs + normalized outputs).
• Develop the test suite with a complete Rust project skeleton first.
• Develop subsystem-by-subsystem until all goldens pass and expanding coverage until parity is reached.

Questions for maintainers:

• Would this be welcome as an official multi-language implementation or preferred as a community project first?
• If official, should it live in this repo (subdir/workspace) or in a separate repo?

Describe alternatives you've considered

• Rust wrapper around Python Burr (PyO3 / embedding Python) to reuse existing behavior but with Rust entry-points.
• Running Burr as a separate Python service and integrating over HTTP/IPC from Rust (keeps Burr unchanged, but adds runtime/ops overhead).
• Partial Rust port (core runtime only) while keeping the existing tracking server/UI and/or persistence layer in Python.
• Implementing only server/API compatibility in Rust while leaving runtime execution to Python (or vice versa).
• Using other Rust state-machine/workflow crates instead of Burr (but that forfeits Burr’s model, tooling, and existing ecosystem).

[elijahbenizzy]

I would be very excited about that -- this is the type of thing that's lower hanging fruit with AI (but still quite an endeavor).

I think same repo -- and might be worth documenting a few APIs happy to review. FWIW I have a typescript implementation in the works. Decided not to go for full parity as the type system capabilities allow for a more powerful API.

Let us know what we can do to support you here!