Zed

Use one Rust-native editor across macOS, Linux, and Windows

The official README describes Zed as a high-performance multiplayer code editor and distributes it for macOS, Linux, and Windows. Rust is the repository's primary language, while a web version is not currently available. The architecture is therefore most relevant when evaluating a desktop-native editing experience rather than a browser IDE.

Sources: [1]

Treat multiplayer editing as a core editor workflow instead of a separate collaboration layer

The collaboration documentation states that multiple people can edit the same project simultaneously and see each other's cursors and edits in real time. Channels provide persistent team rooms with shared projects and voice chat, while contacts and private calls support more ad-hoc sessions.

Sources: [2]

Bring native agents, external agents, MCP, and parallel threads into the editor

Zed's AI documentation provides multiple agent paths: Zed Agent, ACP-integrated External Agents, and terminal-backed threads. Model access can use Zed-hosted models, provider APIs, subscription sign-in, gateways, or local models, while the native agent can also use MCP servers, skills, and instructions. The Threads Sidebar can organize multiple agent threads across projects and worktrees.

Sources: [3]

A strong fit when fast native editing, live collaboration, and agentic development should share one workspace

Zed is worth evaluating when a team wants everyday code editing, pair-style collaboration, and AI-assisted implementation in the same working surface. It is also a useful alternative for developers exploring a Rust-native editor rather than a VS Code-derived stack, or teams that want to operate several agent threads against separate projects or worktrees.

Check platform limits, collaboration access, AI data paths, and mixed licensing separately

The README currently lists macOS, Linux, and Windows, with web support unavailable. Collaboration requires sign-in, and the official docs warn that sharing a project gives collaborators access to the local file system within that project. Teams should therefore scope shared projects and collaborators deliberately.

AI behavior and data boundaries depend on whether you choose Zed-hosted models, a provider API, subscription access, a gateway, or a local model. Licensing is also mixed: the README says Zed source is primarily GPL-3.0-or-later with Apache-2.0 components where marked, so redistribution and reuse should follow the license of the specific code involved.

Sources: [1][2][3]