ZhengyiLuo

Zhengyi Luo is an individual robotics/AI researcher and solo developer who created AgentsDock, an open-source IDE/workspace for agentic AI research. AgentsDock provides a desktop and mobile control surface for Claude Code, Codex, Cursor and long-running research work, with execution handled by a self-hosted AgentsServer on the user's own machine.

Visit Website

At a Glance

2026Est.

Development Environments

Academic Research

Connect

AI Tools by ZhengyiLuo

(1)

AgentsDock

Open Source IDE for AI Agents

AI Coding Asst.Dev Environments Academic Research

Discussions

No discussions yet

Be the first to start a discussion about ZhengyiLuo

Latest News

09/14/2026

AgentsDock 1.0.0 desktop stable released, adding the stable-channel feature set including Cursor, Team Network collaboration, browser chat sharing, scheduled-work improvements and a more resilient multi-server workspace.

github.com

09/14/2026

AgentsDock repository licensed under Apache License 2.0.

github.com

09/13/2026

AgentsDock 1.0.0-beta.2 released with quiet Team Network Mail and Bulletin indicators, explicit refresh behavior and cross-chat timeline regression coverage.

github.com

09/13/2026

AgentsDock 1.0.0-beta.1 released as the first public-source migration beta, moving desktop updates to ZhengyiLuo/AgentsDock and documenting migration compatibility for existing beta installations.

github.com

Products & Services

AgentsDock

2026-09-14 (desktop 1.0.0 stable; beta releases began 2026-09-13)

Open-source desktop and mobile IDE/workspace for agentic AI research. It supports Claude Code, Codex and Cursor, connects to one or more self-hosted servers, and lets users run or resume agent chats, inspect files and rich media, use terminals, schedule work and coordinate agents.

AgentsServer

The self-hosted execution backend paired with AgentsDock. It runs on a workstation or remote machine, exposes authenticated HTTP/WebSocket APIs, streams Claude/Codex events, serves files and videos, manages scheduled jobs and persistent tmux terminals, and keeps project data on the user's own infrastructure.

Market Position

AgentsDock differentiates itself as a self-hosted, multi-device orchestration layer for agent CLIs rather than a single-provider chat UI or a conventional terminal-only SSH client. Its closest alternatives include provider-native Claude Code/Codex/Cursor workflows, terminal/remote-access tools such as Termius, and other agent-development environments; AgentsDock combines their agent access with persistent server-side workspaces, mobile access, file/media review and multi-agent coordination.

Leadership

Founders

Zhengyi "Zen" Luo

Senior Research Scientist at NVIDIA GEAR Lab and former Carnegie Mellon University Robotics Institute PhD researcher. His public biography lists a University of Pennsylvania computer-science bachelor's degree, a CMU robotics PhD, an Apple AR/VR software-engineering internship, NVIDIA Toronto AI Lab and GEAR Lab research internships, and a visiting-researcher role at Meta Reality Labs. He owns the ZhengyiLuo GitHub organization/repositories and created AgentsDock.

Founding Story

Luo describes AgentsDock as a workspace built around his own agentic research workflow: it lets researchers monitor and steer coding, robotics and model-training work from desktop or mobile devices instead of needing to remain at a laptop or terminal. The public product is designed around a client/server split, with private work and credentials remaining on a machine controlled by the user.

Business Model

Revenue Model

Public materials present AgentsDock as open-source software with a self-hosted backend. The app and AgentsServer are distributed through public repositories/releases and user-operated infrastructure; no subscription, usage-based API or commercial licensing model is described.

No IPO or public-market plans are described; this is presented as an individual developer's open-source project.

Target Markets

Industries & Segments

AI researchers
Robotics and humanoid-robotics researchers
Software developers using coding agents
Research labs and technical teams running long-lived jobs
Individual developers who need remote, self-hosted agent access

Use Cases

AI and robotics research workflows
Coding with Claude Code, Codex or Cursor from desktop or mobile
Monitoring model training and other long-running jobs remotely
Reviewing generated plots, images, videos, files and code changes
Persistent terminal-based development on a remote workstation
Multi-agent collaboration, delegation and cross-chat context handoffs

Quick Facts

Founded

2026

History & Milestones

2026-09-13

AgentsDock 1.0.0-beta.1 became the first public-source migration beta in the ZhengyiLuo/AgentsDock repository; the notes describe chat sharing, history, mailbox, storage-recovery, UI and navigation changes.

2026-09-13

AgentsDock 1.0.0-beta.2 added quiet Team Network Mail and Bulletin indicators over a metadata-only connection, with explicit refresh and cross-chat timeline regression coverage.

2026-09-14

AgentsDock 1.0.0 desktop stable released, bringing the desktop beta line into the stable channel with Cursor support, richer agent collaboration/Team Network, browser chat sharing, scheduled work and multi-server workspace improvements.

2026-09-14

The AgentsDock repository was licensed under Apache License 2.0.

Key Capabilities

Desktop and mobile workspace spanning macOS, Linux, Windows, iPhone, iPad and Android distributions

Claude Code, Codex and Cursor support, with provider/model discovery and provider-specific commands

Self-hosted authenticated AgentsServer backend; chats, files, credentials and API keys stay on the user's server

Persistent tmux terminal access from any connected device

Scheduled interval and recurring jobs for unattended long-running work

Browse, preview, edit and download code, files, images and videos in a chat workspace

Integrations & Partnerships

Platform Integrations

Claude Code CLI
OpenAI Codex CLI
Cursor agent backend
AgentsServer over authenticated HTTP/WebSocket
tmux persistent terminal sessions
Tailscale private networking
macOS, Linux, Windows, iPhone, iPad and Android distribution channels

Key Partnerships

AgentsDock integrates with the Claude Code and Codex CLI ecosystems and uses users' own provider accounts on the self-hosted server.

The setup documentation recommends Tailscale for secure remote access outside the local network.

The website displays Carnegie Mellon University, UC Berkeley and NVIDIA logos under 'Trusted by researchers at'; it does not describe formal commercial partnerships.

Connect

Website

agentsdock.net

GitHub

ZhengyiLuo

Discord

ZGDrhEWqPt

AI Topics

ZhengyiLuo focuses on these topics:

AI Coding Assistants(1)

Development Environments(1)

Academic Research(1)

Back to all developers Suggest an edit

ZhengyiLuo