Developers

Build on AssetScope

A REST API to assess telemetry, read predictions, and drive governed remediation — with API-key auth, an OpenAPI spec, and signed webhooks.

OpenAPI spec Get an API key

Authentication

Every request carries an API key issued from the admin portal, scoped per organization:

Authorization: Bearer ak_…

Scopes: read, assess, ingest, remediate, admin. Requests are rate-limited per key.

Connectors

Read-only by construction: AssetScope reads telemetry and history, and never writes to a control system, historian or DCS.

Live telemetry

Prometheus / NVIDIA DCGM · AMD ROCm · AWS Neuron · Google TPU · OSIsoft/AVEVA PI Web API · Modbus TCP · OpenMetrics push · canonical batch API

Historian & CM exports

Process-historian CSV (PI, PHD, Exaquantum) · machinery-protection / CM-route exports · OPC-UA / MQTT / Modbus / BACnet tag dictionaries · SCADA, J1939, IEC 61850, HL7 device metrics

CMMS / EAM

SAP PM and IBM Maximo: history and BOM import; draft notifications / work orders created only on approval, never released · ServiceNow · Jira

For your agents

MCP server (POST /api/mcp): tenant-scoped, read-only tools your own agents can call · Server-Sent Events stream (GET /api/v1/stream) · signed webhooks

Endpoints

GET	/api/v1/health	Service + key health
POST	/api/v1/telemetry/ingest/batch	Canonical batch telemetry ingestion (scope: ingest)
GET	/api/v1/assets/{id}/risk-assessment	Full risk assessment for one equipment item (scope: read)
GET	/api/v1/stream	Server-Sent Events: predictions and alerts as they change (scope: read)
POST	/api/v1/pilot/snapshot	Record today's prediction snapshot for the pilot validation report (scope: assess)
GET	/api/v1/pilot-report	Audit-ready pilot validation report (scope: read)
POST	/api/v1/assess	Score telemetry with the reliability engine
POST	/api/v1/spectral	Vibration analysis for bearing/gear fault detection (BPFO/BPFI/BSF/FTF)
POST	/api/v1/plan	Prescriptive maintenance plan — optimized service schedule from current predictions under crew/budget/horizon constraints
POST	/api/v1/inspect	AI visual/thermal inspection — structured equipment condition report from an image
GET	/api/v1/predictions	List incident predictions for the key's organization
GET	/api/v1/remediation	List remediation requests
POST	/api/v1/remediation	Request a remediation (requires 'remediate' scope; starts PENDING approval)
POST	/api/v1/remediation/{id}	Decide/execute a remediation: action = approve \| reject \| execute (requires 'remediate' scope)

Example — assess telemetry

curl -X POST https://assetscope.ai/api/v1/assess \
  -H "Authorization: Bearer ak_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "assetId": "gpu-1", "assetType": "gpu",
    "samples": [
      {"metricName":"gpu_temperature_c","value":88,"timestampMs":1730000000000},
      {"metricName":"corrected_ecc_errors","value":140,"timestampMs":1730000000000}
    ]
  }'

Returns failure probability (7/30-day), remaining useful life with a calibrated prediction interval, root cause, and a recommended action.

Webhooks

Subscribe to prediction.created, alert.fired, and remediation.executed. Each delivery is signed so you can verify authenticity:

X-AssetScope-Event: remediation.executed
X-AssetScope-Signature: sha256=<HMAC_SHA256(secret, rawBody)>

Governed remediation

Remediation is separation-of-duties by default: POST /remediation creates a pending request, which must be approved and then executed (dry-run unless the executor's integration is configured). Every execution emits a signed remediation.executed webhook and is recorded in the tamper-evident audit log.