Releases·Tool Release
Show HN: Put an AI agent on a FaceTime audio/video call (open source, WebRTC)
Chert (YC P26) released an open-source framework for placing AI agents on FaceTime audio and video calls via WebRTC, plus a managed inbound/outbound API. Because FaceTime has no open API and isn't AppleScript-scriptable, the framework drives the FaceTime UI programmatically and routes agent audio/video through physical loopback machines. The bet is that video agents will replace VoIP as a dominant human-agent channel by meeting users on trusted, familiar platforms.
1 engagement1 sourceWed, Sep 16, 2026, 03:41 PM
AI Overview
- Chert (YC P26) open-sourced a framework putting AI agents on FaceTime audio/video calls via WebRTC, with code at github.com/cherthq/chert-facetime-opensource.
- Chert drives the FaceTime UI programmatically and routes agent audio/video through physical loopback machines, since FaceTime has no open API or AppleScript support.
- Chert offers a managed inbound/outbound API and live demo at trychert.com/facetime, plus a BYOR (Bring Your Own Room) feature.
- Chert argues FaceTime's personal, familiar channel suits video agents better than embedded widgets or Zoom/Meet links used by current video agent tools.
Related
Releases·Tool Release
Agentcall lets coding agents join Zoom/Meet/Teams as live participants
Agentcall is a new tool that allows coding agents like Cursor, Claude Code, Codex, and Gemini CLI to join Google Meet, Zoom, or Teams calls as real participants. They can talk, listen, see the screen, and present slides live, running locally without sending call audio/video/transcripts to the cloud.
Agentcall is a new tool that allows coding agents like Cursor, Claude Code, Codex, and Gemini CLI to join Google Meet, Zoom, or Teams calls as real participants. They can talk, listen, see the screen, and present slides live, running locally without sending call audio/video/transcripts to the cloud.
5 engagement1 sourceredditFri, Jul 17, 2026, 03:38 AM
Products·launched
Free on-device voice control for AI agents with voice reply
This is a free, on-device application that enables voice interaction with AI coding or general agents, allowing users to speak commands and hear responses. It solves the problem of hands-free agent control and can also connect meeting recordings to agents or send agents to participate in meetings actively.
This is a free, on-device application that enables voice interaction with AI coding or general agents, allowing users to speak commands and hear responses. It solves the problem of hands-free agent control and can also connect meeting recordings to agents or send agents to participate in meetings actively.
5 engagement1 sourceredditThu, Jul 30, 2026, 02:45 PM
Products·launched
Privacy-first AI agent browser side panel for meetings, notes, and chats
ChatPanel is a browser side panel that integrates AI agents to assist with meetings, note-taking, and chat. It uses LLMs locally or via bring-your-own-agent (Codex, Claude, etc.) with privacy features like NER and PII redaction. It solves the problem of having an AI that understands your work context while keeping data private.
ChatPanel is a browser side panel that integrates AI agents to assist with meetings, note-taking, and chat. It uses LLMs locally or via bring-your-own-agent (Codex, Claude, etc.) with privacy features like NER and PII redaction. It solves the problem of having an AI that understands your work context while keeping data private.
2 engagement1 sourcehackernewsTue, Jul 21, 2026, 11:56 AM
Social·Community
Why voice and messaging channels beat desktop dashboards for running autonomous agents
A prompt2bot post argues agent frameworks over-index on browser dashboards and traces, which suit development but not daily operation. The authors built Mentat around phone calls, WhatsApp group chats, and an independent cloud environment, treating the agent as an asynchronous collaborator rather than something supervised in real time.
A prompt2bot post argues agent frameworks over-index on browser dashboards and traces, which suit development but not daily operation. The authors built Mentat around phone calls, WhatsApp group chats, and an independent cloud environment, treating the agent as an asynchronous collaborator rather than something supervised in real time.
5 engagement1 sourceredditThu, Sep 10, 2026, 10:30 PM
Products·building
AutoYou — phone app to voice call, chat with, and remote-view your own PC's local AI
AutoYou is a companion app that lets you use your phone to talk to, chat with, and control an AI running on your own computer. It runs speech recognition, transcription, and voice playback locally so calls and chats stay private. It targets people who want multi-modal AI access on the go without routing data through corporate servers.
AutoYou is a companion app that lets you use your phone to talk to, chat with, and control an AI running on your own computer. It runs speech recognition, transcription, and voice playback locally so calls and chats stay private. It targets people who want multi-modal AI access on the go without routing data through corporate servers.
3 engagement1 sourceredditWed, Sep 9, 2026, 01:37 AM