Quickstart: your first workflow

This walks through creating an account, building a simple workflow, and running it — both from the visual canvas and via the API.

1. Register & sign in

Open the frontend (http://localhost:5173) and register. The first registration on a fresh database creates:

• A tenant (your workspace), identified by a tenant_slug
• A user with the owner role on that tenant, on the Free plan

Or via the API:

curl -X POST http://localhost:3001/api/v1/auth/register \
  -H 'Content-Type: application/json' \
  -d '{
    "tenant_slug": "acme",
    "tenant_name": "Acme Inc",
    "email": "you@acme.com",
    "password": "supersecret123"
  }'

This returns a JWT — pass it as Authorization: Bearer <token> on every subsequent API call.

2. Build a workflow on the canvas

1. Click New Workflow from the dashboard.
2. Drag a trigger node onto the canvas — for this example, choose Webhook.
3. Add an HTTP node, connect it to the trigger, and configure it to call any API (e.g. a public JSON API).
4. Add an LLM node, connect it after the HTTP node, and write a prompt that references the HTTP response, e.g. Summarize this data: {{steps.http_1.output.body}}.
5. Add a Slack connector node (or Set node if you don't have Slack connected) to do something with the result.
6. Click Save, then Enable the workflow.

See Workflows & the DSL for how the canvas maps to the underlying JSON definition, and Nodes for what every node type does.

3. The same workflow as DSL

Every workflow is a JSON document with a trigger and a list of steps (a DAG via depends_on). The canvas is just an editor for this document — you can also create workflows directly via the API or the AI Workflow Generator.

{
  "id": "wf_quickstart",
  "name": "Fetch and Summarize",
  "description": "Fetch JSON from an API and summarize it with an LLM",
  "version": 1,
  "trigger": {
    "type": "http",
    "config": {}
  },
  "steps": [
    {
      "id": "fetch",
      "name": "Fetch data",
      "type": "http",
      "config": {
        "method": "GET",
        "url": "https://api.example.com/data"
      }
    },
    {
      "id": "summarize",
      "name": "Summarize with LLM",
      "type": "llm",
      "depends_on": ["fetch"],
      "config": {
        "provider": "anthropic",
        "model": "claude-sonnet-4-6",
        "prompt": "Summarize this JSON in 3 bullet points:\n\n{{steps.fetch.output.body}}"
      }
    }
  ]
}

Create it via the API:

curl -X POST http://localhost:3001/api/v1/workflows \
  -H "Authorization: Bearer $TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "Fetch and Summarize",
    "description": "Fetch JSON from an API and summarize it with an LLM",
    "dsl": { "...": "the JSON above" }
  }'

4. Run it

Manually, via the dashboard's Run button, or:

curl -X POST http://localhost:3001/api/v1/workflows/$WORKFLOW_ID/trigger \
  -H "Authorization: Bearer $TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{}'

Via webhook (if the trigger type is http), enable the workflow and POST to:

curl -X POST http://localhost:3001/webhooks/$WORKFLOW_ID \
  -H 'Content-Type: application/json' \
  -d '{"hello": "world"}'

Each run creates a Temporal workflow execution. Open the Executions tab (or the Temporal Web UI at http://localhost:8081) to watch it progress step by step, inspect each step's input/output, token usage, and cost.

5. Generate a workflow with AI instead

Skip steps 2-3 entirely — describe what you want in plain English:

curl -X POST http://localhost:3001/api/v1/workflows/generate \
  -H "Authorization: Bearer $TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"prompt": "Every morning at 9am, fetch the top 5 Hacker News stories and post a summary to Slack"}'

This returns a complete WorkflowDSL (trigger, steps, connectors, prompts) ready to save and enable. See AI Workflow Generator.

Next steps

• Core Concepts — understand workflows, nodes, triggers, and executions in depth.
• Connector Reference — connect Slack, Gmail, GitHub, databases, and 60+ other services.
• AI Agents & Tools — build agents that can use tools and remember context.