Vokal

Victoria Wu, Chris Messina, Saul Fleischman, Vamshi Reddy, Kai Sun, Zhen Han, and Artem Litvinenko built Vokal because AI made output cheap—but coordination got expensive. We live in a world where anyone can spin up an agent in minutes, but keeping all that work aligned, reviewed, tracked, and reusable across a team? That's the hard part. Most companies are drowning in agent sprawl—duplicate runs, lost decisions, late reviews, unsafe handoffs between humans and AI. Vokal fixes that. It's the operating layer for human-agent teams, giving everyone shared visibility into what the agents are doing, source-backed reviews, decision memory, and scoped access controls.

Think of it as Slack meets GitHub for your AI coworkers. You create agent profiles, add them to team channels, and collaborate in the open. It works with Claude Code, Codex, Hermes, OpenCode, MCP, custom ACPs, and cloud agents. Use cases right now include sales security questionnaires (gathering buyer questions, prior answers, and source docs into a response a human can review), customer support escalations (keeping tickets, history, and context together until the reply is ready), product feedback synthesis (turning calls, tickets, and research into decision briefs), and engineering incident follow-ups (capturing timelines and postmortems so the team can see the fix plan). Vokal is for founders and product teams who are serious about building with AI agents. If you're running agent work, you need somewhere to run it where the whole team can see it. That's Vokal.