API Documentation

The AI Cloud exposes two REST APIs: one for managing AI models and catalogues, and one for managing edge devices and sites. Both are documented via OpenAPI (Swagger UI).

API	OpenAPI UI	Description
Compute API	api.sclbl.nxvms.com/cpt/openapi.html	Model upload, conversion, catalogue management
Device API	api.sclbl.nxvms.com/dev/openapi.html	Site, server, and edge device management

The raw OpenAPI JSON specifications are available at /cpt/openapi.json and /dev/openapi.json on the same host.

The Swagger UI at api.sclbl.nxvms.com/cpt/openapi.html lists all available Compute API endpoints

---

Authentication

All API endpoints require a Bearer token obtained via the Nx OAuth2 flow.

Getting a token

The token exchange happens automatically when you log in through the AI Cloud admin interface. For programmatic access, follow the OAuth2 authorization code flow against the Nx Cloud identity provider:

# Exchange an authorization code for tokens
curl -X POST https://cloud.networkoptix.com/cdb/oauth2/token \
  -H "Content-Type: application/x-www-form-urlencoded" \
  -d "grant_type=authorization_code&code=<your-code>&redirect_uri=<your-redirect-uri>&client_id=<your-client-id>"

The response includes an access_token (JWT). Use it as the Bearer token in all subsequent API calls.

Note: The code parameter is obtained via the OAuth2 authorization code redirect flow, which requires a browser. For interactive testing, use the Swagger UI linked in the Authentication section above. Headless programmatic authentication is not yet covered here.

Using the token

Pass the token in the Authorization header on every request:

export TOKEN="eyJ..."

# List all models available to your account
curl https://api.sclbl.nxvms.com/cpt/functions \
  -H "Authorization: Bearer $TOKEN"

# List all sites
curl https://api.sclbl.nxvms.com/dev/systems \
  -H "Authorization: Bearer $TOKEN"

---

Multi-tenant context (acting as an organization)

Users with access to multiple organizations can make API calls on behalf of a specific organization by passing two additional headers:

Header	Values	Description
X-Act-As-Kind	user, organization, channel_partner	The entity type to act as
X-Act-As-ID	UUID of the organization	The organization UUID

# List models as a specific organization
curl https://api.sclbl.nxvms.com/cpt/functions \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-Act-As-Kind: organization" \
  -H "X-Act-As-ID: <org-uuid>"

Omitting these headers causes the API to act as the authenticated user (personal context).

---

Pagination

List endpoints return paginated results. The total record count is exposed in a response header:

Header	Description
X-Pagination-Records	Total number of records matching the query (regardless of page size)

Pagination parameters vary by endpoint; see the OpenAPI spec for limit and offset query parameters.

# Fetch the second page of models (20 per page)
curl "https://api.sclbl.nxvms.com/cpt/functions?limit=20&offset=20" \
  -H "Authorization: Bearer $TOKEN" \
  -v 2>&1 | grep "X-Pagination-Records"

---

Common operations

Upload a model

curl -X POST https://api.sclbl.nxvms.com/cpt/functions \
  -H "Authorization: Bearer $TOKEN" \
  -F "file=@my-model.onnx" \
  -F "name=My Object Detector" \
  -F "documentation=Detects objects in video streams."

Get a specific model

curl https://api.sclbl.nxvms.com/cpt/function/<model-uuid> \
  -H "Authorization: Bearer $TOKEN"

List devices for a site

curl https://api.sclbl.nxvms.com/dev/system/<site-nxid>/devices \
  -H "Authorization: Bearer $TOKEN"

Assign a model pipeline to a device

curl -X POST https://api.sclbl.nxvms.com/dev/system/<site-nxid>/device/<device-nxid>/functions \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"FunctionID": "<model-uuid>"}'

---

For a full list of endpoints, request/response schemas, and query parameters, use the interactive OpenAPI documentation linked at the top of this page.