8719Inference service port
0.89Ambiguity tolerance
900sPressure window
Docker pull
Registry Command
Use the latest VANTH runtime image for deployment validation. Production deployments should pin a digest once release channels are finalized.
docker pull synantix/vanth-engine:latest
docker run --rm \
--name vanth-inference \
-e VANTH_COGNITIVE_MODE=pressure-based \
-e VANTH_PRESSURE_WINDOW=900s \
-e VANTH_INFERENCE_DEPTH=layered \
-e VANTH_AMBIGUITY_TOLERANCE=0.89 \
-v /var/synantix/telemetry/input:/opt/vanth/telemetry/input:ro \
-v /var/synantix/telemetry/output:/opt/vanth/telemetry/output \
-p 8719:8719 \
synantix/vanth-engine:latestProduction-Shaped Compose File
version: "3.9"
services:
vanth-inference:
image: synantix/vanth-engine:latest
container_name: vanth-inference
restart: unless-stopped
environment:
VANTH_COGNITIVE_MODE: "pressure-based"
VANTH_PRESSURE_WINDOW: "900s"
VANTH_INFERENCE_DEPTH: "layered"
VANTH_AMBIGUITY_TOLERANCE: "0.89"
VANTH_OUTPUT_FORMAT: "structured-json"
volumes:
- /var/synantix/telemetry/input:/opt/vanth/telemetry/input:ro
- /var/synantix/telemetry/output:/opt/vanth/telemetry/output
- /var/synantix/vanth/cache:/opt/vanth/cache
ports:
- "8719:8719"
healthcheck:
test: ["CMD", "vanthctl", "health", "--json"]
interval: 30s
timeout: 5s
retries: 3
Health response
Runtime Verification
{
"service": "vanth-inference",
"instrument": "VANTH_Inference_Core",
"status": "ready",
"cognitive_mode": "pressure-based",
"pressure_window": "900s",
"ambiguity_tolerance": 0.89,
"telemetry_mounts": {
"input": "read-only",
"output": "write-enabled"
}
}
Next step
Architecture Review
Use the review page to model telemetry paths, ambiguity thresholds, and deployment boundaries before connecting VANTH to critical systems.