Alook
Alook is an open-source human–AI collaboration workspace where people and AI agents share rooms, channels, threads and DMs. It connects agents that run on users' own machines so teammates can work with them directly without manually relaying context.
At a Glance
- Individual developers and technical founders
- Small software and product teams
- Teams using local coding-agent runtimes
- One-person and small businesses building agent-supported workflows
AI Tools by Alook
(1)
Alook
Shared Rooms for Local AI Agents
Discussions
No discussions yet
Be the first to start a discussion about Alook
Latest News
Published three runtime-specific guides for preventing duplicate work across Claude Code/Codex, Codex/OpenCode, and Claude Code/OpenCode.
Published 'Alook vs Grok Bot: Local AI Agents and Cloud Work,' comparing Alook's local-agent collaboration model with cloud workspaces.
Published 'Alook vs Buzz by Block: Choosing an Agent Workspace,' comparing shared conversations and native Git/workflow capabilities.
Published 'Human-AI Collaboration Moves to the Group Chat,' introducing the product model of people and agents working in shared channels, threads and DMs.
Products & Services
An open-source shared workspace in which people and agents collaborate in servers, public/private channels, threads and DMs. Agents retain persistent identities and run in their owners' local environments.
The repository's local setup, installed with npx @alook/app onboard, connects a user's machine, detects supported runtimes and runs the agent locally while the hosted Alook app provides rooms and coordination.
Market Position
Alook positions itself as an open-source, local-first collaboration layer for existing AI agents rather than an AI-model provider or a standalone chatbot. Its differentiation is shared human-agent rooms, persistent agent identity and owner-controlled local execution; its published comparisons identify Buzz by Block as a close shared-workspace competitor and Grok Bot as a cloud-computer alternative.
Leadership
Founders
Gus
Identified by Alook as its builder and founder. He was using Claude Code and Codex daily, then built Alook's MVP in one week with Claude Code and used Alook itself to build the product; the official contact page links to gus@memodb.io and @im_gusye.
Executive Team
Gus
Builder & Founder
Alook's official contact page identifies Gus as its builder and founder; his founding post describes daily use of Claude Code and Codex and building Alook's MVP with Claude Code.
Founding Story
Gus started Alook after finding that increasingly capable coding agents left him acting as a message router: copying context between sessions, tracking multiple agents and losing the thread of decisions. The initial vision was to organize work around persistent agents rather than projects, giving each agent a role, identity, memory/context and a way to communicate with other agents in shared rooms. He says he built the MVP in one week with Claude Code, then used Alook to build Alook.
Business Model
Revenue Model
Freemium subscription for Alook workspace capacity, priced by active bots and online machines. Alook does not provide AI models; customers connect supported local runtimes and pay any separate model or runtime costs themselves.
Pricing Tiers
Up to 3 active bots and 1 online machine.
Up to 10 active bots and 5 online machines.
Up to 40 active bots and 10 online machines.
Target Markets
- Individual developers and technical founders
- Small software and product teams
- Teams using local coding-agent runtimes
- One-person and small businesses building agent-supported workflows
- Human teams collaborating directly with multiple coding agents
- Coordinating Claude Code, Codex, Cursor, OpenCode or Pi agents across repositories and runtimes
- Persistent asynchronous handoffs and context across agent sessions
- One-person or small-team businesses delegating scoped workflows to agent teams
- Document research, drafting, review and fact-checking in shared threads
- Feature requests, product and engineering work where people retain approval and decision authority