Quick Start

Get your first annotated image in minutes. This guide walks you through the Playground flow: upload → prompt → process → result.

Before You Start

Sign in — Open the Playground and sign in with your account. You need to be signed in to create tasks and use credits.
Consent — The first time you run a task, you may be asked to accept image processing terms. Accept to continue.
Credits — Your plan (Free, Starter, or Pro) provides a credit balance. Each annotation uses credits based on the model and resolution you choose.

Go to the Playground (link in the nav or Getting Started).
In the main card (Image & Prompt), add an image:
- Click the image area or the + button and choose a file, or
- Drag and drop an image onto the card, or
- Paste from the clipboard (e.g. a screenshot).
Supported formats: PNG, JPEG, WebP.

You must upload at least one image before you can process. You can remove and replace it anytime before submitting.

In the same card, use the prompt field to tell the AI what to focus on or how to annotate.
Leave it empty to use the default annotation style (the system will apply a built-in template).
Use Use Template (when the field is empty) to fill in a sample prompt, then edit if you like.
You can type or paste markdown; the prompt is optional.

Model — Pick an AI model (e.g. Nano Banana Pro) from the dropdown. Availability depends on your plan.
Resolution — For models that support it (e.g. Nano Banana Pro, Nano Banana 2), choose output resolution (e.g. 1K, 2K, 4K). Higher resolution may use more credits and can be gated by plan.
Web Search (Starter/Pro) — If available, you can enable web search so the AI can use up-to-date context in the annotation.

The UI shows an estimated credit cost for the current image count and options. Ensure you have enough credits before submitting.

Click Submit (or press Ctrl+Enter / Cmd+Enter) to submit the task.
The app will:
- Create a task and check your credits and consent.
- Upload your image(s) to the server.
- Submit the job to the AI pipeline.
The view switches to the task/chat view: your task appears in the sidebar and the main area shows the task status and reference image(s).

Processing usually takes 10–30 seconds, often under a minute, depending on image complexity.

In the task view you'll see:
- Reference image(s) — What you uploaded.
- Job status — Pending → Queued → Processing → Completed (or Failed).
- Annotated image(s) — Once the job completes, the annotated result is shown.
You can download or copy the annotated image and use it elsewhere.
From here you can also Edit prompt & Rerun to change the prompt or model and run again on the same image(s).

Step	What you do
Upload	Add at least one image (file, drag-and-drop, or paste).
Prompt	Optionally enter or template a prompt; leave empty for default.
Process	Choose model/resolution, then click Submit (or Ctrl/Cmd+Enter).
Result	Watch status in the task view; download or copy the annotated image when done.