Chert

Chertとは？

Chert is a FaceTime-focused AI video agent product that lets teams configure an assistant, assign it to a managed line, and publish it for controlled calling workflows. The page positions it as “Vapi for FaceTime,” with a focus on agents that can use camera context during calls instead of relying only on voice.

It is designed for situations where seeing the problem matters as much as hearing it. The source highlights remote support, field service, telehealth intake, guided onboarding, visual inspections, and customer success as examples of work that can benefit from a video agent. The product also emphasizes a controlled private preview, so the FaceTime offering is not described as generally available yet.

Chertでできること

Camera-aware call handling

The product is built around the idea that an agent can use what the caller is showing on camera to understand a problem or next step more directly than voice alone.

Prompt and behavior configuration

Users can define the assistant’s instructions, realtime model, voice, and behavior before a call is placed.

Persona and framing selection

The setup flow includes choosing an avatar persona and framing style, such as a mid-shot presentation.

Browser preview before publishing

A browser-based preview is available to check the prompt, voice, microphone behavior, interruption handling, avatar, and cleanup before going live.

Managed line deployment

The assistant can be deployed by publishing it and assigning it to a provisioned FaceTime line.

Inbound and outbound test workflows

The control plane supports bounded inbound and outbound test workflows, with live execution described as provisioned and explicitly authorized.

利用シーン

“Remote support”

Guide a caller through a problem they can show on camera, such as a cable, device, or environment issue.

“Field service assistance”

Help users or technicians walk through a physical setup or repair step by step when visual confirmation is useful.

“Telehealth intake”

Use a video agent for intake workflows where a person may need to show symptoms, equipment, or surroundings.

“Guided onboarding”

Walk new users through a process visually, reducing the need for them to describe every detail from memory.

“Visual inspections and customer success”

Support review workflows where the agent needs to see an object, location, or setup in order to proceed.

よくある質問

Is the FaceTime product generally available?

No. The page says the FaceTime product is being onboarded through a controlled private preview while production media and line readiness are validated.

Can I try it without placing a live FaceTime call?

Yes. The browser preview is meant for checking the prompt, voice, microphone, interruption behavior, avatar, and cleanup before publishing.

Does it support both inbound and outbound calls?

The page says the control plane supports bounded inbound and outbound test workflows. Live execution is provisioned and explicitly authorized, and accepting mode is not enabled by default.

What can be configured in the assistant?

The page says you can configure the assistant instructions, realtime model, voice, avatar, framing, publishing state, and managed line assignment.

What kinds of workflows is it meant for?

The product is presented for situations where a caller needs to show a problem or environment, including remote support, field service, telehealth intake, guided onboarding, visual inspections, and customer success.