Guides

Voice calls and character voices

Updated October 18, 2026

On Shiro you can call a character and talk out loud, like a phone call. They answer in their own voice, react to what you say, and everything said in the call is added to your chat, so the story carries on in text afterward. This guide explains how calls work, how call minutes work, how character voices are chosen, and how to get the most natural conversation.

Starting a call

Open a chat and tap the call button at the top. The phone rings and the character picks up with a short line, the way they would answer. From then on, just talk. Shiro listens for when you stop speaking, sends what you said, and plays the character's reply. You don't need to press anything between turns.

Your browser or phone will ask for microphone access the first time. Calls work in the browser on phones and computers, and in the Android app.

Talking naturally

Why a reply sometimes takes a moment

Each turn has three steps: turning your speech into text, writing the character's reply, and turning the reply into speech. Shiro does these as quickly as it can and starts playing the first sentence of a reply while the rest is still being voiced, so longer replies don't make you wait longer. If your connection drops for a moment, the app retries on its own.

Call minutes

Calls use call minutes. Only talking counts: the time you speak and the time the character speaks, not the silence in between. There are a few ways to get minutes:

The call screen shows how much time you have left. When it runs out, the call ends and the app shows ways to get more.

How character voices are chosen

Every character has a voice. When a character is created, Shiro picks one that suits them, from a range of voices with different ages, tones and styles, based on the character's gender, personality and description. It also tries to give different characters different voices, so your favorites don't all sound alike.

Creators can also choose a voice themselves. When you create or edit a character, open the Voice section, filter by gender, and tap a voice to hear a sample. Pick one, or leave it on Automatic to let Shiro choose. The same voice is used for calls and for voice replies.

Performed in the moment

Voices aren't read flatly. Each reply is performed with the character's personality and the mood of the moment in mind: a teasing line sounds teasing, a worried one sounds worried. Actions in *asterisks* aren't read aloud, only what the character says.

Voice replies in text chats

With shiro+, you can also have replies in a normal text chat read aloud in the character's voice. If you've been listening to replies in a chat, new ones are voiced as they arrive so they're ready when you tap play.

Privacy

Your voice is turned into text to write the reply, and the text is saved in your chat like any message. Like all messages on Shiro, it's checked against our community guidelines. See the Privacy Policy for details.

Troubleshooting

Open Shiro

More guides