We built a way for Claude Code to join meetings like a real teammate
Hi HN — we built agentcall.dev because the coding agent you're already running in your terminal shouldn't be trapped there. The pitch: your existing Claude Code, Codex, OpenClaw, or Cursor session joins a Google Meet, Teams, or Zoom call as itself. Same session, same context, same file access. It speaks, listens, screen-shares a localhost webpage, and can code live while you all talk about what it's building. What's actually on the call: • Voice in, voice out. Two modes — collaborative (sub-second via a voice intelligence layer tuned for latency) or direct (~2s, your coding agent itself…
In plain words
AgentCall.dev enables coding agents like Claude Code to participate in video calls on Google Meet, Teams, or Zoom as active participants. The agent joins with its existing session context and file access, speaking and listening in real-time while sharing a localhost webpage and writing code during discussions. It offers two communication modes: a collaborative option with sub-second latency or a direct mode where the agent handles its own responses with full reasoning displayed. The screen-sharing feature exposes only what the agent chooses via a secure temporary tunnel, preventing unintended information leakage. It is built for teams using coding agents who want their AI participants fully integrated into meetings.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
Hi HN — we built agentcall.dev because the coding agent you're already running in your terminal shouldn't be trapped there. The pitch: your existing Claude Code, Codex, OpenClaw, or Cursor session joins a Google Meet, Teams, or Zoom call as itself. Same session, same context, same file access. It speaks, listens, screen-shares a localhost webpage, and can code live while you all talk about what it's building. What's actually on the call: • Voice in, voice out. Two modes — collaborative (sub-second via a voice intelligence layer tuned for latency) or direct (~2s, your coding agent itself doing the talking, with full reasoning). • Screen share that is not a desktop grab. It's a URL or local port rendered into the meeting through a per-call tunnel (temp subdomain + secret token, dies when the call ends). The agent chooses what to expose, nothing else leaks. • Shareable webpage links that participants open in their own browser — live dashboards, diffs, forms — served from the agent's localhost, tunneled only for the call. • Meeting chat: the agent reads incoming messages (great for dropping in URLs, error logs, code snippets that sound terrible over TTS) and sends them back — your agent can paste a PR link while still talking. • Participant awareness: the agent knows who's in the room, who joined, who left, and who's actively speaking. It can raise its hand to speak politely, toggle its own mic, and change voice/interruption/barge-in settings on the fly. • Leave and rejoin on command. Ask mid-call ("leave for two minutes and come back") and the agent does it. • TS/STT run on our own servers (several voices, English STT today, multi-language TTS). No data to third parties in direct mode. Everything above is controlled by your coding agent through the skill. There are no hardcoded meeting behaviors — the agent decides when to speak, when to chat, when to raise a hand, what to share. What we don't do: we don't run the coding agent. The agent's model calls, file I/O, tokens, and tools all happen on your machine or your cloud. We don't store recordings or screen captures by default — everything streams to the agent in real time and the agent decides what to keep. Transcripts are in-memory for crash resilience and wiped on disconnect unless you opt into retention (up to 7 days). We bundle TTS + voice + tunneling, start at $0.35/hour (drops with volume), and store nothing by default. We join existing meetings — we're not a new video platform. Get AgentCall Running — 2 Minutes Works with Claude Code (best), Codex, and Cursor. 1. Install the skill install join-meeting skill from https://github.com/pattern-ai-labs/agentcall 2. Get your API key Sign up at https://agentcall.dev , copy your key from the dashboard, and paste it when the skill asks. 3. Invite the agent to a meeting Paste any Google Meet / Zoom / Teams URL into your agent: Give it 30–60 seconds for the bot to spin up. Once it joins, start talking. Demo: https://www.youtube.com/@pattern-ai-labs Would love feedback on the voice latency in direct mode, the privacy model, and where the "joining existing meetings" framing breaks for your workflow.
Does the same job
all alternatives →
Claude Code Desktop App RedesignedApr 2026 · ▲565Run parallel coding agents from one desktop workspace
- ASAgentBox – SDK to Run Claude Code, Codex, or OpenCode in Any SandboxApr 2026 · github.com · ▲8
Hi HN, AgentBox is an SDK for running coding agents (Claude Code, Codex, OpenCode) inside sandboxes (Docker, E2B, Modal, Daytona, Vercel). One API. Swap the agent or the sandbox and your code doesn't change. Think of it as what the AI SDK did for LLMs, but for agent + runtime. Most wrappers call agents in non interactive mode (claude --print, codex exec). AgentBox instead boots each agent's native server inside the sandbox (Codex app-server over JSON-RPC, OpenCode serve over HTTP/SSE, Claude Code's SDK WebSocket transport) and drives it from the host, behaving like an interactive…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, April 2026
the whole month →
- AG
Thought the resources for GPU arch were lacking, so here we are
Life & fun · Apr 2026 · jaso1024.com
- IB
Built a ~9M param LLM from scratch to understand how they actually work. Vanilla transformer, 60K synthetic conversations, ~130 lines of PyTorch. Trains in 5 min on a free Colab T4. The fish thinks the meaning of life is food. Fork it and swap the personality for your own character.
AI · Apr 2026 · github.com
- BC
Life & fun · Apr 2026 · sam-burns.com
- IB
With social media and now AI, its important to keep the indie web alive. There are many people who write frequently. Blogosphere tries to highlight them by fetching the recent posts from personal blogs across many categories. There are two versions: Minimal (HN-inspired, fast, static): https://text.blogosphere.app/ Non-minimal: https://blogosphere.app/ If you don't find your blog (or your favorite ones), please add them. I will review and approve it.
AI · Apr 2026 · text.blogosphere.app