Real-Time Streaming Decode
Callback-based streaming decode at <100ms latency. Configurable epoch length and overlap. Thread-safe command output queue.
Developer SDK
Integrate motor-imagery decoding in hours, not months.
Python and C++ SDKs with a REST API layer. Three integration paths for every deployment context — from prototype to production clinical device.
SDK capabilities
Everything in the decode pipeline, exposed.
Clean, typed API surface with minimal boilerplate. You own the EEG hardware layer — we handle the signal science.
Batch Offline Analysis
Process BDF/GDF recordings offline. REST API endpoint for cloud batch jobs. Returns class labels, confidence scores, and decision time series.
Session-to-Session Alignment
Euclidean alignment applied automatically on session initialization. Maintains classification accuracy across sessions without retraining. Zero-config default.
Command Output Formats
Digital I/O (TTL), serial UART (ASCII or binary), UDP packet, Python callback. One session configuration, multiple simultaneous output streams.
Calibration Wizard
Guided calibration sequence: motor-imagery instruction prompts, real-time epoch quality feedback, CSP filter training, baseline normalization. 5–10 min per session.
Logging & Telemetry
Structured JSON session logs: decode events, confidence scores, latency measurements, artifact rejection counts. Local file or REST endpoint target.
Python SDK
Start decoding in 8 lines.
The Python SDK wraps the full decode pipeline. Import, connect your EEG source, calibrate, then stream. The C++ SDK exposes identical semantics for embedded device integration.
Python SDK — pip install synaptiq. Requires Python ≥3.9, numpy, scipy.
C++ SDK — CMake package. C++17, headers-only decode API. Linux & Windows.
REST API — POST/GET JSON over HTTPS. Cloud batch + remote session status.
integrate_sdk.py
import synaptiq
from synaptiq.output import UDPOutput
# Connect to EEG via Lab Streaming Layer
session = synaptiq.Session(
device="lsl",
channels=16,
srate=500
)
# Run calibration sequence (8 min)
session.calibrate(classes=["left_hand", "right_hand"])
# Route commands to UDP + Python callback
session.add_output(UDPOutput(host="192.168.1.100", port=5005))
def on_command(cmd):
print(f"[{cmd.latency_ms:.1f}ms] {cmd.label} ({cmd.confidence:.2f})")
# Start real-time decode — blocks until session.stop()
session.decode_stream(callback=on_command)Request SDK access.
We review each request to ensure device compatibility and clinical context. Development tier is free. Clinical and Enterprise tiers start at $2,400/device/year.
Request SDK Access