Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.synheart.ai/llms.txt

Use this file to discover all available pages before exploring further.

This guide takes you from a fresh machine to a working Synheart integration. Pick the track that matches what you’re building.

​
Pick a track

Track A — Synheart Core (HSI)

Multimodal human-state inference on-device. Requires CLI, platform account, and the runtime.

Track B — Standalone module

Just synheart_wear, synheart_behavior, or synheart_session. No CLI required.

​
Track A

Synheart Core (HSI). Use this track when you want fused human-state output (focus, emotion, capacity, recovery, strain, sleep) from biosignals + behavior + phone context.

​
1. Install the CLI

curl -fsSL https://synheart.sh/install | sh
Windows (PowerShell): 
iwr -useb https://synheart.sh/install.ps1 | iex
Verify: 
synheart version
Full install guide →

​
2. Create a platform account

Synheart is invite-only. Request access, then create your account at platform.synheart.ai (email + password or OAuth). Account setup →

​
3. Authenticate the CLI

synheart login
Follow the device-flow prompt to sign in. Confirm: 
synheart whoami
Authentication details →

​
4. Install the runtime

From your project root: 
synheart install runtime
This downloads the native HSI engine and writes synheart.lock. Commit the lockfile so teammates and CI can run synheart sync. Runtime install guide →

​
5. Initialize Synheart Core

# pubspec.yaml
dependencies:
  synheart_core: ^0.0.4

import 'package:synheart_core/synheart_core.dart';

await Synheart.initialize(
  userId: 'anon_user_123',
  config: SynheartConfig(
    allowUnsignedCapabilities: true, // use capabilityToken in production
  ),
);

Synheart.onHSIUpdate.listen((hsiJson) {
  print('HSI: $hsiJson');
});

// build.gradle.kts
dependencies {
    implementation("ai.synheart:synheart-core:0.0.4")
}

import ai.synheart.core.Synheart
import ai.synheart.core.config.SynheartConfig

Synheart.initialize(
    context = context,
    userId = "anon_user_123",
    config = SynheartConfig(allowUnsignedCapabilities = true),
)

Synheart.onHSIUpdate.collect { hsiJson ->
    println("HSI: $hsiJson")
}

// Package.swift
.package(url: "https://github.com/synheart-ai/synheart-core-swift.git", from: "0.0.4")

import SynheartCore
import Combine

var cancellables = Set<AnyCancellable>()

try await Synheart.initialize(
    userId: "anon_user_123",
    config: SynheartConfig(allowUnsignedCapabilities: true)
)

Synheart.onHSIUpdate
    .sink { hsiJson in print("HSI: \(hsiJson)") }
    .store(in: &cancellables)

​
6. Test locally with mock data (optional)

synheart local        # mock platform server (consent + ingest APIs)
synheart mock start   # mock wearable stream on ws://localhost:8787
CLI reference →

​
Track B

Standalone module. If you only need a single signal channel and don’t need fused HSI output, install just that SDK. No CLI, account, or runtime needed.

Synheart Wear

Biosignals from Apple Watch, WHOOP, Garmin, Fitbit, BLE HRMs.

Synheart Behavior

Tap, scroll, type, and motion patterns — content-free.

Synheart Session

Time-bounded sessions with HR + behavioral metrics on-device.
Quick example — synheart_wear in Flutter: 
dependencies:
  synheart_wear: ^0.3.1

import 'package:synheart_wear/synheart_wear.dart';

final synheart = SynheartWear(
  config: SynheartWearConfig.withAdapters({DeviceAdapter.platformHealth}),
);
await synheart.initialize();

synheart.streamHR(interval: const Duration(seconds: 5)).listen((m) {
  print('HR: ${m.getMetric(MetricType.hr)}');
});

​
How it fits together

  1. CLI authenticates you, manages the runtime artifact, and runs local mocks for development.
  2. Runtime is the native on-device engine that computes HSI. The Core SDK loads it at startup.
  3. Core SDK (Flutter / Kotlin / Swift) coordinates signal collection, drives the runtime, and exposes reactive streams.
  4. Modules (synheart_wear, synheart_behavior, synheart_session) can be used standalone or composed under Core.

​
Privacy posture

On-device

Inference runs locally — no biosignal content leaves the device.

Consent-gated

Cloud uploads require explicit user consent and capability tokens.

Content-free

No text, audio, screen content, or PII is ever captured.

​
Next

Synheart Core

Architecture, HSI/HSV spec, capabilities, consent.

CLI reference

Mock streams, local platform, install, sync, receiver.

Modules

Wear, Behavior, Session details.

Troubleshooting

Common issues and fixes.