Give tasks from the bottom input

Tell the Agent what to make from the bottom input, point it at nodes, and add files.

The input at the bottom of the canvas is where you give the Agent work. Hold Space to say it, or press / to type it.

1. Type a task

  1. Press /, or click the input.
  2. Type what you want.
  3. Press Enter to send.

Typing a task in the bottom input

Copy this instruction and try it with TapNow Agent

Make a 15-second teaser for a new oat-milk latte. Vertical 9:16, morning café mood, warm light. First write four shots, then make one key image per shot.

You can keep talking while the Agent works. New messages join the work in progress.

2. Point the Agent at nodes

To tell the Agent which nodes to use:

  • Select them before you send. They show as small thumbnails on the input.
  • Or type @ and pick a node from the list.

When you use several nodes, say what each one is for:

Copy this instruction and try it with TapNow Agent

Use @Bottle Front to keep the label and shape exactly. Use @Sunset Beach only for the light and colors. Make two 4:5 product shots.

3. Use a skill

Skills give the Agent a special ability, such as a house style or a ready-made workflow. Type / at the start of your message, pick a skill from the list, then write your request.

The skills list

Find more skills in the App Market.

4. Add a document

To have the Agent read a brief, PDF, or Word file, drag it onto the input. Or click + and choose Attachment. The file is used for this message only.

To keep an image, video, or sound on the canvas instead, use Upload. See Add files and assets.

5. Check that your message arrived

After you send, a small note above the input shows which Agent picked up your message.

The note above the input after sending

If it says Not delivered, click Resend.

Next: Talk with your voice