---
title: "Voice calls and phone numbers in Chat Thing"
canonical_url: "https://chatthing.ai/blog/voice-calls-and-phone-numbers"
last_updated: "2026-09-30T15:52:50.838Z"
meta:
description: "Talk to your Chat Thing agent from your web chat, or give it a US or UK phone number. Calls use shared 60-minute voice packs."
"og:description": "Talk to your Chat Thing agent from your web chat, or give it a US or UK phone number. Calls use shared 60-minute voice packs."
"og:title": "Chat Thing agents can talk: voice calls and phone numbers are here - Blog - Chat Thing"
"twitter:description": "Talk to your Chat Thing agent from your web chat, or give it a US or UK phone number. Calls use shared 60-minute voice packs."
"twitter:title": "Chat Thing agents can talk: voice calls and phone numbers are here - Blog - Chat Thing"
---
[← Back to blog](https://chatthing.ai/blog)**Changelog**# **Chat Thing agents can talk: voice calls and phone numbers are here**
Zef
30 Sept 2026
~ 4 min read
---
---
### **On this page**
1. [Voice calls in your web chat](#voice-calls-in-your-web-chat)
2. [A phone number your agent answers](#a-phone-number-your-agent-answers)
3. [Same agent, same model](#same-agent-same-model)
4. [Pricing](#pricing)
5. [Try it](#try-it)
Your Chat Thing agent can now talk.
Visitors can press a button in your web chat and have a conversation out loud. Customers can ring a phone number and your agent picks up. Either way it's the agent you've already built, with the same knowledge, power-ups and AI model, now with a voice.
This has been one of the most enjoyable things we've worked on in a while. Hearing an agent you configured answer a question from your own docs, then change the page you're looking at while it talks, is quite something.
## Voice calls in your web chat
Turn on voice calls and your agent's web chat gets a call button. Visitors allow their microphone, start talking and hear your agent reply in the same window, with live captions if they want them. They can interrupt, change their mind or ask a follow-up without starting again, and when they're done they can carry on typing in the same conversation.
It gets more interesting on a website your agent can operate. If you've given it website power-ups, it can use them during a call. Here's a real clip from our property demo: we ask to see the results on a map, the agent says it's on it, and the page switches from property cards to map markers while the conversation carries on.
_A continuous 14-second clip at normal speed, including the wait. The properties are demo data._
And here's a longer one. Our Skin Routine Finder asks about skin type, concerns and budget, suggests a routine, then adjusts it three times as the shopper changes their mind. When the catalogue can't confirm something, it says so.
What you can set up:
- 13 voices with British, Irish, Australian, North American and other accents, with previews that don't use minutes.
- Voice instructions for personality and tone, drafted from your agent's system prompt and then yours to edit.
- The spoken language, including following the visitor's own language.
- Short, medium or long answers.
- A maximum call length of 1 to 15 minutes, and an optional monthly limit per agent.
Everything said on a call is saved to the conversation, so it's in your inbox like any other chat.
Find it under your agent's **Web channel → Advanced features → Voice calls**. For the full walkthrough, including the demos above, read [Introducing voice calls](https://chatthing.ai/blog/introducing-voice-calls).
## A phone number your agent answers
The new **Phone** channel gives your agent a phone line. Rent a US local or UK mobile number, or forward the number your customers already know, whether that's every call, only after hours or only when you're busy.
On a call, your agent:
- Answers with your greeting and a short line saying it's an AI assistant.
- Answers questions from your website, docs and files, and can use server-side power-ups while it talks: look up an order, send an email, create a ticket or book a meeting through Cal.com.
- Transfers callers to your team when they ask for a person. The person answering presses a key to accept, so voicemail can't take the call.
- Takes a message if nobody picks up: name, callback number and what they need, read back to the caller to confirm.
Every call is saved as a text transcript in your inbox, marked with a phone icon. Call audio isn't recorded. Limits are on from the start to protect your minutes: three calls at once, 30 calls per caller per day, withheld numbers rejected and a block list.
You can hear it before spending anything. Turn on the Phone channel, get a PIN for your own number and ring the shared test line. Every team on the Standard plan or above gets 10 free test minutes a month.
_The test line section. The numbers shown are fictional test numbers._
The Phone channel is available on the Standard plan and above. Turn it on from your agent's **Channels** tab, and see the [phone channel docs](https://chatthing.ai/docs/phone-channel) for setup.
## Same agent, same model
Voice calls run on two layers. OpenAI's GPT-Live-1 handles the spoken conversation: listening, speaking and coping with interruptions. When a question needs your business knowledge or someone asks for something to be done, it hands the request to your Chat Thing agent, which does the work with its own model, instructions and power-ups.
So adding voice doesn't mean switching your agent to a different AI model. If it runs on Claude, Gemini or an OpenAI model today, that model keeps doing the thinking, on the website and on the phone.
## Pricing
Voice is a monthly add-on to any paid plan, managed under **Billing → Voice calls**.
| What | Price |
| --- | --- |
| Voice pack: 60 minutes a month, shared across your team's agents and between web and phone calls | $10/month |
| Phone number, US local or UK mobile | $5/month each |
| Phone test line | 10 free minutes a month |
There are no overage charges. When minutes run out, new calls are politely declined and text chat keeps working. The work your agent does during a call uses your normal message tokens, the same as in chat.
To celebrate the launch, eligible teams get $5 off their first voice invoice, so your first 60 minutes cost $5. See [voice calls and phone numbers](https://chatthing.ai/docs/voice-calls) for how minutes are counted.
## Try it
Pick an agent you know well, add a voice pack and turn on voice calls in its web channel. Ask it something from your knowledge base, then a follow-up, then something that needs one of its power-ups. For the phone, start on the free test line.
We'd love to hear what you build with it, and what you'd like voice to do next.
[Open Chat Thing](https://app.chatthing.ai/app/bots)
### **Related articles **
[](https://chatthing.ai/blog/glm-5-3-flash-supportbench-and-link-controls) [GLM 5.3 Flash lands in Chat Thing, plus SupportBench and link controls
](https://chatthing.ai/blog/glm-5-3-flash-supportbench-and-link-controls)
GLM 5.3 Flash is now available in Chat Thing, alongside SupportBench model testing and control over where links in embedded agents open.
[](https://chatthing.ai/blog/gemini-3-7-flash-and-grok-4-6) [Gemini 3.7 Flash and Grok 4.6 land in Chat Thing
](https://chatthing.ai/blog/gemini-3-7-flash-and-grok-4-6)
Gemini 3.7 Flash and Grok 4.6 are now available in Chat Thing, alongside clearer Power-Up activity and support for more Google Drive file types.
[](https://chatthing.ai/blog/ask-otto-and-seven-new-models) [Otto now lives inside every agent, plus seven new models
](https://chatthing.ai/blog/ask-otto-and-seven-new-models)
Otto now lives inside every agent, so you can change things by describing what you want rather than hunting through settings forms. Plus Claude Opus 5 and six more models, four new embed capabilities, and twelve retired models.
- **Previous **
- [← Turn a WebMCP-enabled website into a copilot with Chat Thing](https://chatthing.ai/blog/chat-thing-webmcp-website-copilot)