API access & credentials.

Ordel is built API-first: everything the web app does, it does by calling the same HTTP API you can call yourself. There are three ways to integrate, depending on what you are building.

1. Coding agents over MCP (recommended)

The sanctioned path for an AI agent (Claude, Cursor, or your own) is the Model Context Protocol server. It exposes Ordel's capabilities as typed tools, with scoped OAuth and per-tool permissions, so an agent can plan and run tests without you wiring raw endpoints.

Start here: Connect your coding agent.

2. The REST API with an API key

For scripts, backend integrations, and one-off automation, authenticate with a personal API key.

Create a key: Account → API keys (personal), or Admin → Members → API keys (workspace). Choose a scope and, optionally, an expiry.

• Keys are read-only by default.
• A key that needs to create or change data needs a write scope; a key that triggers runs needs a run scope.
• A key acts as you, with your role's permissions, and only inside the tenant it was issued for. It cannot cross tenants. If a key leaks, revoke it from the same screen.

Call the API: send the key as a bearer token.

curl -s https://ordel.io/api/test-cases/<projectId> \
  -H "Authorization: Bearer $ORDEL_API_KEY"

Every request is scoped to your tenant automatically, so you do not send a tenant id for a single-tenant key. Writes are rejected with 403 unless the key carries the matching scope.

A curated, browsable API reference (endpoints, request/response shapes, "try it") is on the way. Until it ships, the endpoints mirror the app: if the UI can do it, there is an endpoint behind it, and the resource names line up with what you see in the product (requirements, test cases, runs, cycles, scripts, reports).

3. Trigger runs from CI

To kick off a test run from a pipeline, use a run-scoped API key with the CI trigger endpoint. The easiest way in is the GitHub Actions workflow Ordel generates for you: open Applications, pick the app, then the Git tab, and copy it from CI integration → CI YAML. See Run tests on your machine and Agent tokens vs API keys for how credentials differ between the API and a runner.

Not sure which credential you need? Agent tokens vs API keys lays out the difference between authenticating you (an API key) and authenticating a machine (a runner's agent token).