We built it to solve the common problem we and our customers had: engineers spend more time on AI ops and plumbing than on actual feature development. This was often due to the challenge of using Python for I/O-bound, high-concurrency workloads and complexity maintaining pipelines with streams that use always-changing ML models.
Our solution is a high-performance runtime written in C++ with the core idea of defining AI logic as graphs. For instance, a basic voice-to-voice agent consists of STT → LLM → TTS nodes, while the connecting edges stream data and enforce conditions. This graph engine is portable (Linux, Windows, macOS) and can run on-device.
We built a few key features on top of this C++ core:
- Extensions. Runtime architecture decouples graph definition from implementation. If a pre-built component doesn't exist, you can register your own custom node/code and reuse it in any graph without writing any glue code.
- Routers. You can dynamically select models/settings on the per-node basis depending on the traffic as well as configure policies for fallbacks and retries to get the app ready for production.
- The Portal. A web-based control plane UI to deploy graphs, push config changes instantly, run A/B tests on live traffic, and monitor your app with logs, traces, and metrics.
- Unified API. Use our optimized models or route to providers like OpenAI, Anthropic, and Google through a single, consistent interface and one API key.
We have a Node.js SDK out now, with Python, Unity, Unreal, and native C++ coming soon. We plan to open-source the SDKs, starting with Node.js.
The docs are here: https://docs.inworld.ai/docs/runtime/overview
We're eager for feedback from fellow engineers and builders. What do you think?
skyzouwdev•5mo ago
Curious about the choice of C++ for the core runtime. Was the main driver raw performance, or more about portability and low overhead in production environments? In my case, I’ve mostly worked with Python pipelines and they tend to choke when you try to scale concurrency without overcomplicating the architecture.
The graph-based approach reminds me a bit of Unreal’s Blueprints, but for AI ops. If you can make it genuinely easy to plug in custom nodes without touching glue code, that could be a big productivity boost.
rogilop•5mo ago
Originally, Inworld was mostly involved into gaming (read as Unreal & Unity) and therefore the choice. Yet (surprise, surprise) we want native x-platform and on-device (whose official support is coming soon).
Also, not a new thing, but C++ is more reliable than Python under really high concurrency, which was critical for us (a lot of people using same graphs in parallel kind of use case).