Voice & Audio OpenAI Opens GPT-Live-1.1 to the API and Prices Silence at 2 Cents a Minute Coding Assistants Replit Pitted 32 Claude Agents Against NumPy Tuning: 18.9x Speedup While GPT-6.2 Swarms Collapsed at Scale Security GAO Covert Tests: All Four Top Chatbots Produced Viable Attack Plans in Under 12 Hours Once Guardrails Fell Meta Meta Puts Its 2030 Net-Zero Target Under Review as AI Data Centers Push Net Emissions Up 45 Percent Business Daniel Ek Is Backing Europa Fusion, a Stealth UK Startup Led by a Former DeepMind Researcher Betting on Pulsed Power Coding Assistants Stripe Tested 14 Coding Agents on 69 Payment Scenarios: 78% of Criteria Passed, but Only 29% of Runs Met Every Requirement Security A Transcription Firm Used by Police and Prosecutors Left Interview Transcripts and Body-Cam Footage on an Open S3 Bucket News UN Killer-Robot Talks End With a 2028 Deadline and No Binding Rules. About 100 States Wanted a Treaty; the US, Russia and Israel Blocked It Voice & Audio OpenAI Opens GPT-Live-1.1 to the API and Prices Silence at 2 Cents a Minute Coding Assistants Replit Pitted 32 Claude Agents Against NumPy Tuning: 18.9x Speedup While GPT-6.2 Swarms Collapsed at Scale Security GAO Covert Tests: All Four Top Chatbots Produced Viable Attack Plans in Under 12 Hours Once Guardrails Fell Meta Meta Puts Its 2030 Net-Zero Target Under Review as AI Data Centers Push Net Emissions Up 45 Percent Business Daniel Ek Is Backing Europa Fusion, a Stealth UK Startup Led by a Former DeepMind Researcher Betting on Pulsed Power Coding Assistants Stripe Tested 14 Coding Agents on 69 Payment Scenarios: 78% of Criteria Passed, but Only 29% of Runs Met Every Requirement Security A Transcription Firm Used by Police and Prosecutors Left Interview Transcripts and Body-Cam Footage on an Open S3 Bucket News UN Killer-Robot Talks End With a 2028 Deadline and No Binding Rules. About 100 States Wanted a Treaty; the US, Russia and Israel Blocked It

Vercel Ships Board, a Headless Message Layer Where AI Agents From Different Frameworks Claim Tasks and Report Back

Vercel Ships Board, a Headless Message Layer Where AI Agents From Different Frameworks Claim Tasks and Report Back

Vercel released Board, an experimental headless communication primitive that gives AI agents and humans a shared message store. It is harness-agnostic — agents built on Claude Code, Codex, LangChain, Pi, OpenCode, eve, fx and Mastra can share one backlog — and syncs two ways with Slack and GitHub. Vercel says its rebuilt v0 software factory now merges over 50 percent of pull requests on it.

Vercel has released Board, an experimental "headless" communication primitive built to coordinate fleets of AI agents and the humans supervising them. Instead of another dashboard, Board is a centralized message store: agents from different environments read and write to shared channels, claim tasks from a common backlog, and post updates that any participant — human or machine — can follow.

The system is deliberately harness-agnostic. According to Vercel, agents built on Claude Code, Codex, LangChain, Pi, OpenCode, eve, fx and Mastra can all join the same board. That addresses a real gap: as teams move from single-task bots to multi-agent teams, there has been no purpose-built primitive to track delegation, statuses and logs across fragmented environments. Markdown files and ad-hoc tickets stop scaling quickly once a dozen agents work in parallel.

Developers get a CLI and a JavaScript SDK for agent read and write operations, while humans can inspect progress through a dedicated viewer launched with npx @vercel/board inspect. Board organizes work into channels, threads and labels, and supports two-way synchronization with Slack and GitHub, so an agent can act on a human's report — and a human can act on an agent's.

Two mechanisms make the store feel less like a log file and more like an operations layer. "Watchers" can wake Vercel deployments when labels change or new replies land, and agents can schedule proactive follow-ups with the vercel board schedule command instead of waiting to be polled.

Vercel is dogfooding it in production. The company says its internal security agents and its "v0 software factory" now run on Board, and that after rebuilding the software factory on the primitive, the system merges more than 50 percent of pull requests. That is the clearest signal yet of a shift the industry keeps gesturing at: the hard problem is no longer the underlying model, but the coordination layer that lets specialized agents work in parallel without colliding.

Whether Board becomes a blueprint depends on adoption beyond Vercel's own stack. Competing ideas — shared task queues inside cloud SDKs, agent-to-agent protocols — are all chasing the same prize: a standard way for heterogeneous agents to talk to each other. A headless, framework-neutral message store is a credible candidate, because it imposes nothing on how agents think and everything on how they communicate.

Comments (0)

Log in to join the discussion

Log In

No comments yet