Calls
Calls is app-to-app voice. Human, AI and agent-initiated calling have separate service, consent and access requirements.
Access is by request. Windows, macOS, Android, iOS: in development. No public installers or store listings are approved. Check platform details
App-to-app voice is designed for two authorized human participants.
An approved Calls-enabled build, accepted contact and fresh service authorization.
Real-device audio and release approval remain pending. Human calls do not require AI-provider tools. Group call actions select one recipient.
AI voice is a separate service that processes audio and selected context.
Separate AI voice availability, explicit acceptance and disclosed processing.
Provider/privacy and real-speech qualification remain pending. Proposed work requires review; a proposal is not execution or a sent chat message.
Agent-initiated calls have their own availability and consent requirements.
Separate authorized service availability and recipient opt-in.
Outbound speech and quiet-hours qualification remain pending; no public availability is approved.
A person on the other end
The human-call design is for two independently authorized people with an accepted contact relationship. A call action in a group selects one recipient; it does not create a group conference.
Human Calls do not depend on AI-provider tools or an AI voice service.
Invitations, Answer/Decline, mute, disconnect audio, output selection, recents, availability/quiet hours and returning to an active call are implemented in the candidate. Installed audio and lifecycle behavior still need qualification for the approved build.
The app must be running for the documented desktop incoming path. Reliable background or locked-device delivery, cross-device transfer and every OS audio route are not qualified.
Human call media uses the media service. AI calls also send intelligible audio and selected authorized context to the voice worker and AI service. The agent is identified as AI.
A proposed task from AI voice needs review in the supported workspace path. A proposal is not an already executed task or a message already sent to chat.
Transport encryption does not establish end-to-end secrecy from media or AI processors. Disabled application recording does not prove zero provider retention. Provider settings, regions and lifecycle disclosures still need review before release.
Read about data and choicesIn-app calling requires independent OS qualification.
Qualified OS build, explicit consent and approved service availability.
Physical-device audio and lifecycle acceptance are not verified. No reliable background delivery is promised.
Background and locked-device calling requires independent OS qualification.
Qualified OS build, explicit consent and approved service availability.
Physical-device audio and lifecycle acceptance are not verified. No reliable background delivery is promised.
In-app calling requires independent OS qualification.
Qualified OS build, explicit consent and approved service availability.
Physical-device audio and lifecycle acceptance are not verified. No reliable background delivery is promised.
Background and locked-device calling requires independent OS qualification.
Qualified OS build, explicit consent and approved service availability.
Physical-device audio and lifecycle acceptance are not verified. No reliable background delivery is promised.
Video, screen sharing, group conferencing, telephone-network calls, saved recordings, automatic transcripts and voice notes are not offered by this site.
Bring minds together. Move work forward.
Explore human conversations, AI agents and supported work coordination.
Access is by request. Windows, macOS, Android, iOS: in development. No public installers or store listings are approved. Check platform details
