Workflows

All paths below are relative to https://app.innflow.ai/api/v1. Use bearer authentication and an explicit workspace. Reads require workflows:read; writes require workflows:write for scoped keys.

List workflows

GET /workflows returns standard workflows, ordered by most recently updated first.

curl --fail-with-body 'https://app.innflow.ai/api/v1/workflows?page=1&pageSize=20&search=Leasing' \
  -H "Authorization: Bearer $INNFLOW_API_KEY" \
  -H "X-Innflow-Workspace: $INNFLOW_WORKSPACE"

{
  "data": [{
    "id": "workflow_example",
    "name": "Leasing digest",
    "icon": null,
    "iconColor": null,
    "deploymentStatus": "DRAFT",
    "deployedAt": null,
    "createdAt": "2026-09-28T12:00:00.000Z",
    "updatedAt": "2026-09-28T12:00:00.000Z"
  }],
  "pagination": { "page": 1, "pageSize": 20, "totalCount": 1, "totalPages": 1 }
}

Create a workflow

POST /workflows creates a draft with an initial placeholder node. It does not create a complete automation. The optional JSON name sets the title. Returns 201 with the workflow summary under data.

curl --fail-with-body -X POST 'https://app.innflow.ai/api/v1/workflows' \
  -H "Authorization: Bearer $INNFLOW_API_KEY" \
  -H "X-Innflow-Workspace: $INNFLOW_WORKSPACE" \
  -H 'Content-Type: application/json' \
  -H 'Idempotency-Key: create-leasing-digest-001' \
  -d '{"name":"Leasing digest"}'

Configure its trigger, nodes, and connections in the editor, or import a definition. Use a new idempotency key for each intentionally new workflow.

Read a workflow

GET /workflows/{workflowId} returns { "data": ... } with metadata, nodes, and connections. Node entries contain id, type, name, position, and data. Connections use fromNodeId, toNodeId, fromOutput, and toInput.

curl --fail-with-body "https://app.innflow.ai/api/v1/workflows/$WORKFLOW_ID" \
  -H "Authorization: Bearer $INNFLOW_API_KEY" \
  -H "X-Innflow-Workspace: $INNFLOW_WORKSPACE"

Optional view=summary returns metadata and node/connection counts instead of the graph. nodeIds accepts a comma-separated selection of at most 50 node IDs; it filters nodes, not the connections array.

Update metadata

PATCH /workflows/{workflowId} accepts at least one of name, icon, or iconColor. name must be a nonempty string. icon and iconColor can be cleared with null. This endpoint updates metadata, not the workflow graph. Returns 200 with the updated summary under data.

curl --fail-with-body -X PATCH "https://app.innflow.ai/api/v1/workflows/$WORKFLOW_ID" \
  -H "Authorization: Bearer $INNFLOW_API_KEY" \
  -H "X-Innflow-Workspace: $INNFLOW_WORKSPACE" \
  -H 'Content-Type: application/json' \
  -d '{"name":"Daily leasing digest"}'

Import a definition

POST /workflows/import creates a new draft. It accepts a JSON object with these fields:

Each node requires a supported type and position: { "x": 0, "y": 0 }. It can include name and a data object. Each edge requires integer source and target indices into the nodes array. Optional fromOutput and toInput select handles; sourceHandle and targetHandle are accepted aliases.

This is the smallest illustrative draft definition, not a runnable automation:

{
  "name": "Integration draft",
  "canvasLayoutVersion": 1,
  "nodes": [{ "type": "INITIAL", "position": { "x": 0, "y": 0 }, "data": {} }],
  "edges": []
}

Save the definition as workflow.json, then import it:

curl --fail-with-body -X POST 'https://app.innflow.ai/api/v1/workflows/import' \
  -H "Authorization: Bearer $INNFLOW_API_KEY" \
  -H "X-Innflow-Workspace: $INNFLOW_WORKSPACE" \
  -H 'Content-Type: application/json' \
  -H 'Idempotency-Key: import-integration-draft-001' \
  --data-binary @workflow.json

Returns 201 with id, name, deploymentStatus, nodeCount, edgeCount, createdAt, and updatedAt under data. The import creates fresh node IDs. The read endpoint's connections array is not the import endpoint's edges array: convert node-ID references to array indices before a REST round trip. The CLI provides import/export commands.

The accepted description field is not currently persisted by this import handler. Layout metadata should be preserved across exports and imports. Review connected-account references and node configuration before deploying an imported workflow.

Duplicate a workflow

POST /workflows/{workflowId}/duplicate accepts an optional name. Returns 201 with a new draft summary under data; omitted names use the original name with (Copy). Node IDs are regenerated and layout metadata is preserved. This request supports Idempotency-Key.

Deploy or undeploy

No JSON body is required for these API-key requests. Both return 200 with a workflow summary under data. Deploying an already deployed workflow or undeploying one that is not deployed returns 422.

curl --fail-with-body -X POST "https://app.innflow.ai/api/v1/workflows/$WORKFLOW_ID/deploy" \
  -H "Authorization: Bearer $INNFLOW_API_KEY" \
  -H "X-Innflow-Workspace: $INNFLOW_WORKSPACE"

Deployment can activate configured schedules. Inspect the workflow first. These endpoints do not provide the Idempotency-Key replay behavior used by create/import/duplicate/execute. After an ambiguous timeout, read the workflow's deployment state before retrying.

To start a run, see executions.

Delete a workflow

DELETE /workflows/{workflowId} deletes the workflow and its associated execution history. It returns 200 with:

{ "data": { "id": "workflow_example", "deleted": true } }

Deletion is destructive. Read and confirm the workflow ID and workspace before issuing the request. A resource outside the selected scope returns 404.