Run Your First Workflow in 60 Seconds

See Jetty in action immediately with a simple text-to-image workflow.

Step 1: Copy this workflow

{
  "init_params": {
    "prompt": "A cute corgi wearing a tiny top hat"
  },
  "step_configs": {
    "generate": {
      "model": "black-forest-labs/flux-schnell",
      "activity": "replicate_text2image",
      "prompt_path": "init_params.prompt"
    }
  },
  "steps": ["generate"]
}

Step 2: Run it

Option A: Use the Jetty UI

1. Go to dock.jetty.io
2. Create or select a collection
3. Click "New Task" and paste the JSON
4. Click "Run"

Option B: Use the API

curl -X POST "https://flows-api.jetty.io/api/v1/run-sync/your-collection/your-task" \
  -H "Authorization: Bearer $JETTY_API_TOKEN" \
  -F "bakery_host=https://dock.jetty.io" \
  -F 'init_params={"prompt": "A cute corgi wearing a tiny top hat"}'

Step 3: See your result

The workflow generates an image from your text prompt. That's it!

---

What just happened?

Component	Purpose
init_params	Your inputs (the prompt)
step_configs	What to do (generate an image using Flux)
steps	Execution order (just one step here)

---

How Jetty workflows work

init_params → step_1 → step_2 → step_3 → outputs
                ↓         ↑
            (path expressions connect steps)

Every workflow follows this pattern:

1. Input parameters define what goes in
2. Steps process the data sequentially
3. Path expressions like init_params.prompt connect everything

---

Next: Build something useful

I want to...	Tutorial	Time
Compare AI models side-by-side	Model Comparison	5 min
Generate and evaluate images	Image Generation	5 min
Use LLM-as-Judge for quality scoring	LLM Evaluation	5 min
Process documents with AI	Document Processing	5 min
Run batch evaluations	Batch Processing	10 min
Benchmark AI agents	Agent Benchmarking	10 min

---

Need to set up your environment first?

If you need API tokens or want to understand the full setup:

→ Complete Setup Guide