Roadmap

Phases 1–5 are the showcase roadmap and map one-to-one onto the GitHub milestones. Phases 6–9 come from the sequencing section of the architecture plan, on a horizon of roughly a decade.

Issue state snapshotted 5 August 2026 · live milestones

  1. Phase 1 complete

    Scripting foundation

    8 of 8 issues closed

    The C# side of the engine: lifecycle, math, time, queries, and the error isolation that keeps a bad script from taking the editor down.

  2. Phase 2 in progress

    External projects and hot reload

    4 of 8 issues closed

    Turning the engine into something you can point at your own project: external game assemblies, hot reload, project structure and CLI tooling.

  3. Phase 3 open

    Editor usability and asset pipeline

    0 of 5 issues closed

    The editor becomes the primary authoring tool: script attachment, play/stop mode, scene persistence, and a proper UUID-based asset pipeline.

  4. Phase 4 in progress

    Visual and gameplay foundations

    2 of 7 issues closed

  5. Phase 5 in progress

    Shippability

    1 of 8 issues closed

    Closing the gap between tech demo and shippable product: materials, multi-scene, export packaging, a demo game and documentation. Contributions open after this phase.

  6. Phase 6 planned

    A lot more engine features.

    First half of 2027

    Orrin at this stage is missing a lot of features that are required for games to be made. This phase is trying to tackle this.

    • Render graph inspector and one-file frame capture
    • Shader graph and scripting editing.
    • MCP tooling.
    • Asset hot reloading.
    • Skeletal animation, designed into the ECS before more systems assume static meshes
    • A physics decision, realistically integrating Rapier rather than growing the Phase 1 collision system
  7. Phase 7 planned

    Packages, plugins and assets

    Through 2027

    Cargo's model applied to game content: a package manager over content-addressed assets, self-hostable registries, and a shared import cache. Scheduled before collaboration because it exercises the same server infrastructure at lower protocol risk.

  8. Phase 8 planned

    Real-time collaboration

    2027 into 2028

    The CRDT scene document behind the existing mutation pipeline, the self-hosted server, presence and asset locks — then hosted convenience once self-hosting is stable.

  9. Phase 9 planned

    Real-time path tracing

    Onward

    An alternative render graph over the same scene and materials, benchmarked against the reference tracer that will have been validating the renderer for two years by then.

Contributions open after Phase 5

The project is deliberately closed to contributions until then — the stated reason is that it's early enough that the direction could get misread. Until it opens, discussion happens on Discord.

Join the Discord