Book a build call
Back to Help
Guide

Choose a look for your agent’s face

4 min read

A look is a way of drawing your agent's face. Changing it changes nothing else: it is the same agent, with the same voice, the same knowledge and the same conversation. Only the drawing changes.

Your agent is voice-first and hands-free. It speaks first, in NetShow's own voice, and the visitor's microphone opens only after the visitor taps. Whatever look is showing, the voice carries on.

Where the looks are

Every agent's public page has a small menu button, Avatar menu. Open it and the first row that matters here is Look. Each choice is a radio button; the one that is filled in is the look showing now.

The same menu also holds Voice, and Chat with an Open Chat button, for anyone with no microphone or speaker who would rather type.

The looks

These are the rows the Look menu offers today. An agent shows the ones that have been made for it, so your agent may show fewer.

  • ASCII: the face drawn in text characters that move as it talks.
  • Sketchline: a living ink line drawing of the face.
  • Sketchline color: the same line drawing, in color.
  • SVG: a drawn vector character with a body, which can point with its hand.
  • Hybrid: a living face drawn over the agent's own art or photo; the eyes, brows and mouth move while the picture stays.
  • More looks: Kinetic Typography: the agent's own words become the face.
  • More looks: Glyph eyes: the agent as two simple eyes on a small grid, the kind of face a watch or a small screen can show.

An agent that has no made face yet still gets the looks that need nothing prepared in advance. They are drawn in the browser from the agent's name: ASCII, Sketchline, SVG, Kinetic Typography and Glyph eyes. The menu then says More looks appear once this agent has a made face.

How a visitor switches the look

  1. Press Avatar menu.
  2. Under Look, pick a row.
  3. The face redraws in place. The voice and the conversation carry on.

The choice lasts for this visit only. Nothing is saved to the visitor's account or sent to you, and the next visit opens on your look again. If a look cannot load on that device, the face stays as it was and the menu goes back to the look that was showing.

How you set the look your agent opens with

Your agent opens on the face it is wearing now. That face is chosen when you create the agent: the steps are in Create an agent, and Change how it looks explains the difference between the face and the still picture on cards.

If your agent was made with more than one face, its card on your dashboard shows a Face row of small tiles, one per face, with the current one pressed. Press another tile to switch. The agent's public page then opens on that face, and the filled-in Look row follows it. Switching back is one more press, so there is nothing to confirm.

When you open your own agent's page while signed in, the menu shows you every row, including the ones not made yet, with a short note such as This look has not been made for this agent yet. Keep the current look or choose another. Visitors never see those notes; they see only the looks that are ready.

Slow phones and saved data

Under the face there is a second Look control that starts on Automatic. It picks how heavy a drawing the device can carry:

  • With data saving turned on, it starts on a lighter drawing: Sketch (ASCII) when the agent has one.
  • On a small screen, it starts on Flat face (2D), or a lighter drawing if there is no flat face.
  • A strong enough device can show Full body (3D) when the agent has a 3D body.
  • When the device should not animate the face, for example with reduced motion turned on, it shows a still picture and says Showing a still picture on this device.

A visitor can change this control too. Choices that the agent has no body for are greyed out.

Still stuck?

The longer documentation covers models, voice, knowledge, and workflows in depth.

Open the documentation