Chert
AI video agent that answers and places FaceTime calls with a few lines of code and understands the caller's camera view.
About Chert
Chert provides AI video agents that can answer and place FaceTime calls with a few lines of code. When an issue is visual, the agent can use the caller's camera feed to understand what they are showing, for cases where callers need to show the problem itself, their environment, or the next step. The product is currently being opened gradually through a controlled private preview, while production media and communication-line readiness are still being validated; the console supports bounded inbound and outbound testing flows, and live execution requires configuration and explicit authorization, with accept mode off by default. Before launch, use the browser preview to check prompts, voice, microphone, interruption behavior, avatar, and ending performance, and configure assistant instructions, realtime model, voice, avatar, framing, release status, and hosted line allocation.
Best for
Best use cases
When callers need to show a fault or the problem itself instead of describing it from memory
When the caller's environment or on-site situation needs to be understood through the camera
When callers need guidance through the next step
Features
FaceTime video agent
Answer and place FaceTime calls with a few lines of code.
Camera view understanding
Understand what the caller is showing using camera context.
Browser preview
Check prompts, voice, microphone, interruption behavior, avatar, and ending performance before launch.
Flexible configuration
Configure assistant instructions, realtime model, voice, avatar, framing, release status, and hosted line allocation.
Pros
- Integrate in a few lines of code, with support for answering and placing FaceTime calls
- Use the caller's camera feed to understand what they show, suited to issues rooted in the visual layer
- Browser preview to check prompts, voice, microphone, interruption behavior, avatar, and ending performance before launch
- Rich configuration for assistant instructions, realtime model, voice, avatar, framing, release status, and hosted line allocation
Cons
- The product is currently being opened gradually through a controlled private preview, while production media and communication-line readiness are still being validated
- Accept mode is off by default; live execution requires configuration and explicit authorization