Xybrid

Native SDKs are powered by the Rust runtime and expose its native model capabilities. The Browser/Web preview is a separate browser-native LiteRT.js adapter that consumes model_metadata.json and currently supports only LiteRT raw typed-tensor I/O.

The native SDKs share the Rust core; the Browser/Web preview does not and has the limited LiteRT tensor surface described above.

Install and run a model in your language of choice. Each section includes the install snippet and a minimal example.

See the full Installation Guide for all options.

Install in pubspec.yaml:

dependencies:
  xybrid_flutter: ^0.4.1

Run a model:

final model = await Xybrid.model('kokoro-82m').load();
final result = await model.run(XybridEnvelope.text('Hello world'));
// result → 24kHz WAV audio

Install in build.gradle.kts:

dependencies {
    implementation("ai.xybrid:xybrid-kotlin:0.4.1")
}

Run a model:

val model = Xybrid.model("kokoro-82m").load()
val result = model.runAsync(Envelope.text("Hello world"))
// result → 24kHz WAV audio

Install in Package.swift:

dependencies: [
    .package(url: "https://github.com/xybrid-ai/xybrid.git", from: "0.4.1")
]

Run a model:

let model = try await Xybrid.model("kokoro-82m").load()
let result = try await model.runAsync(envelope: Envelope.text("Hello world"))
// result → 24kHz WAV audio

Install via OpenUPM — openupm add ai.xybrid.sdk, or add the https://package.openupm.com scoped registry for scope ai.xybrid — or straight from the git subfolder:

https://github.com/xybrid-ai/xybrid.git?path=/bindings/unity

Native libraries download automatically on import. See the Unity SDK README for details.

Run a model:

var model = XybridClient.LoadModel("kokoro-82m");
var result = model.Run(Envelope.Text("Hello world"));
// result → 24kHz WAV audio

Install in Cargo.toml:

[dependencies]
xybrid = "0.4.1"

Run a model:

let model = Xybrid::model("kokoro-82m").load()?;
let result = model.run(&Envelope::text("Hello world"))?;
// result → 24kHz WAV audio

Install:

# macOS / Linux
curl -sSL https://raw.githubusercontent.com/xybrid-ai/xybrid/master/install.sh | sh

# Windows (PowerShell)
irm https://raw.githubusercontent.com/xybrid-ai/xybrid/master/install.ps1 | iex

Run a model:

xybrid run --model kokoro-82m --input-text "Hello world" -o output.wav

For platform-specific setup, see each SDK's README: Flutter · Unity · Swift · Kotlin · Rust.

# voice-assistant.yaml
name: voice-assistant
stages:
  - model: whisper-tiny-ggml  # Speech → text
  - model: qwen2.5-0.5b       # Process with LLM
  - model: kokoro-82m         # Text → speech

xybrid run --config voice-assistant.yaml --input-audio question.wav -o response.wav

final pipeline = Xybrid.pipeline(yaml: yamlString);
final result = await pipeline.run(XybridEnvelope.audio(bytes: audioBytes, sampleRate: 16000));

// Multi-model pipeline (MMP) support coming soon — use single model loading for now

// Multi-model pipeline (MMP) support coming soon — use single model loading for now

// Multi-model pipeline (MMP) support coming soon — use single model loading for now

let pipeline = Xybrid::pipeline(&yaml_string).load()?;
pipeline.load_models()?;
let result = pipeline.run(&Envelope::audio(audio_bytes))?;

All models run entirely on-device. No cloud, no API keys required. Browse the full registry with xybrid models list.

# Install xybrid skills into your project
curl -sSL https://raw.githubusercontent.com/xybrid-ai/xybrid/master/tools/scripts/install-skills.sh | sh

# Generate model_metadata.json from a HuggingFace model
claude /xybrid-init hexgrad/Kokoro-82M-v1.0-ONNX

{
  "model_id": "my-model",
  "version": "1.0",
  "execution_template": { "type": "Onnx", "model_file": "model.onnx" },
  "preprocessing": [],
  "postprocessing": [],
  "files": ["model.onnx"],
  "metadata": { "task": "text-generation" }
}

SDK MMP support: Flutter ✅ · Rust ✅ · Kotlin 🔜 · Swift 🔜 · Unity 🔜

Tool calling: local models call functions you define — your tools are plain data (Tool::function(...)) and the loop is your code, so any tooling plugs in. On-device via llama.cpp (LFM2 and gemma-4 protocols); Rust SDK and CLI today (xybrid repl ships built-in web_search + your own via --tools-file), Swift/Kotlin/Flutter bindings next. See the Tool Calling guide.

• Privacy first — All inference runs on-device. Your data never leaves the device. The SDK attaches a small fleet-attribution header on registry metadata calls — see registry telemetry.
• Offline capable — No internet required after initial model download.
• Cross-platform — One API across iOS, Android, macOS, Linux, and Windows.
• Multi-model pipelines (MMP) — Chain models together (ASR → LLM → TTS) in a single call.
• Hardware acceleration — Apple Neural Engine, Metal, CUDA, plus opt-in Vulkan for Linux llama.cpp builds.

• Documentation
• Discord
• X (Twitter)
• GitHub Issues

We welcome contributions! See CONTRIBUTING.md for guidelines on setting up your development environment, submitting pull requests, and adding new models.

New here? Browse the good first issue label for small, self-contained tasks. Tasks are also grouped by area: area: core, area: sdk, area: examples, area: bindings, area: tests. Medium-difficulty tasks live under help wanted.

Apache License 2.0 — see LICENSE for details.