Voice on iOS, in early access

A thin iOS client opens a streaming voice session against your myai workspace.

Tap to talk

A push-to-talk session captures your question, sends it to your workspace over a streaming WebSocket, and reads the response back. Mid-utterance cut-offs that surfaced early in voice testing are gone, and a live event log is available in the app for debugging.

Transcripts written back to your workspace

Every voice session writes its transcript to a Work Order in your workspace. You can search and reference what you asked verbally just like a chat thread.

Orb, sound effects, artifact webview

The session UI includes a redesigned orb that shows speech state at a glance, audio cues for turn boundaries, and a webview for opening artifacts the agent surfaces during the conversation.

Early access: the iOS client ships through TestFlight to the design-partner cohort. Reach out if your team would benefit from a voice surface over their context.