FAQ and technical context.

cpp.js is an open-source orchestration system that compiles C++ libraries for web (WebAssembly), iOS, and Android targets from a single configuration. It handles dependency resolution, automatic bindings generation, and cross-platform packaging. It is the open-source foundation and technical proof layer that Celoce's commercial infrastructure extends.

Emscripten+ is Celoce's extended WebAssembly interoperability layer. It extends and re-architects Emscripten's interoperability capabilities, providing improved embind interoperability, C++ object-oriented interoperability, overloaded function support, and enhanced runtime binding workflows. It does not replace Emscripten; it extends it with interoperability enhancements for production workflows.

JSI++ is Celoce's extended React Native runtime interoperability layer (JSI++: extended React Native runtime interoperability layer). It extends and re-architects JSI's interoperability capabilities, providing embind-compatible JSI integration, C++ object interoperability, unified runtime semantics, and reduced bridge inconsistency across platforms.

Celoce Cloud is distributed build execution and artifact caching infrastructure. It executes builds remotely, caches artifacts for reuse, and distributes compiled outputs via CDN. It provides reduced setup overhead and reusable build outputs compared to local Docker-based workflows.

Celoce Enterprise is a self-hosted build control plane for organizations that require full control over their compilation environment and artifacts. It includes internal binary registries, secure compilation pipelines, governance and audit logging, and reproducible builds infrastructure deployed on your own systems.

No. Celoce has zero AI runtime or build dependency. It does not require LLMs or AI models to compile, package, or run native code. The core infrastructure is deterministic systems software, not AI inference.

The relationship is the reverse: AI-agent workflows and AI-based companies can consume Celoce infrastructure. They can use Celoce through the same commercial infrastructure as human developers. They rely on Celoce infrastructure to gain native C++ performance for their cross-platform workflows. With Celoce, they can use C++ libraries through a unified cross-platform build and interoperability layer.

Yes, deterministic. No, no AI dependency. Celoce has zero AI runtime or build-time dependency. The core infrastructure is deterministic systems software: the same inputs always produce the same artifacts. AI agents are consumers of the API, not part of the engine.

Emscripten compiles C/C++ to WebAssembly. cpp.js and Celoce extend Emscripten with workflow abstraction: automatic dependency resolution, bindings generation, multi-target output (web + mobile), artifact caching, orchestration, and deployment consistency.

For single-file compilations, Emscripten alone may be sufficient. For libraries with transitive dependencies and multi-platform targets, the orchestration layer reduces significant operational complexity. See the detailed comparison.

Bazel and Buck2 are build orchestrators optimized for very large monorepos - their unit of work is computing the minimum rebuild set across a graph of targets. Celoce's unit of work is different: compile and package a C/C++ library as a polyglot artifact for web, iOS, Android (and soon Rust, Go, Swift, Kotlin), with bindings generated automatically.

The closer comparables are vcpkg and Conan - package managers for C/C++ - neither of which targets WebAssembly or mobile as first-class outputs. Celoce extends that pattern to the modern target set, with a managed cloud option (Celoce Cloud) and a self-hosted control plane (Celoce Enterprise) for teams that need it.

Yes. cpp.js runs entirely locally using Docker-based toolchains. No account or cloud connection is required. Celoce Cloud is optional and provides remote build execution for teams that want to avoid maintaining local build environments.

Celoce Cloud caches compiled artifacts by content hash. When a build is requested, the system checks whether an identical configuration (same source, same dependencies, same target) has been built before. If a cached artifact exists, it is returned without recompilation. This is particularly effective for libraries that change infrequently but are rebuilt often.

Celoce Cloud executes builds on behalf of customers; it does not itself redistribute compiled outputs. License obligations of underlying libraries - including any LGPL or GPL components - remain with the customer running the build, the same as if they had run the toolchain locally. The Celoce Hub catalog identifies the license of each library so teams can make informed adoption decisions.

For Enterprise customers, we generate compliance artifacts (notice files, SBOM exports, source-availability declarations) as part of the build pipeline, turning licensing rigor into a supported workflow rather than a customer burden.

Compute-heavy workloads see the most significant benefits: geospatial processing, numerical computing, media encoding/decoding, physics simulation, and cryptographic operations. These are domains where C/C++ implementations have years of optimization that JavaScript alternatives cannot easily replicate. See supported workloads for details.

Not all C++ libraries are portable to WebAssembly. Libraries that depend on OS-specific APIs, GPU drivers, or kernel features may not compile. Threading support depends on environment configuration (SharedArrayBuffer, COOP/COEP headers). Performance varies by workload: compute-bound tasks benefit most. See the Known Limitations section below for details.

Minimal. Celoce works with Claude Code, Cursor, Codex CLI, Copilot CLI, Gemini CLI, OpenCode, and any agent that can follow instructions. Human developers can start just as quickly with npm create cpp.js@beta.

Yes. AI-agent workflows can use Celoce through the same APIs, SDKs, and CLI tooling as human developers. They call Celoce Cloud APIs for compilation, trigger distributed build and packaging pipelines, and run end-to-end build workflows.

AI-generated application code is often written in Python, JavaScript, or TypeScript. With Celoce, AI-agent workflows can move performance-sensitive parts into C++ while targeting supported web and mobile platforms. They can add custom C++ code and use native original unmodified C++ libraries. Celoce provides the infrastructure; the AI-agent workflows are the consumers.