AI/TLDR

ZCode

Z.ai's desktop coding agent, tuned for GLM-5.3, that runs multi-agent goals and can be driven from a chat app

Autonomous Coding AgentsCommercial
License
Proprietary (commercial subscription)
Coverage
2 stories

Overview

ZCode is Z.ai's own desktop coding agent, which the product page describes as "next-gen vibe coding for complex goals—with multiple agents and control from anywhere". It is built around GLM-5.3, Z.ai's flagship model, which the company says is tuned for ZCode so that agentic coding runs faster and more steadily. GLM-5.3-Flash is bundled alongside it for multimodal work such as reading a screenshot or analysing an image.

The app is a native download rather than a terminal tool. Z.ai ships installers for macOS on both Apple Silicon and Intel, for Windows on x64 and ARM64, and a beta Linux build as .deb, .rpm and AppImage packages for x64 and ARM64. Work is organised into Goals, the unit ZCode uses for long-running tasks, and a bot-control feature lets you steer a running goal from WeChat, Feishu or Telegram instead of sitting at the machine.

ZCode is a paid subscription with three tiers: Lite at $12.60 a month with 10,000 credits a week, Pro at $56 a month with about 60,000, and Max at $117.60 a month with about 140,000. Every tier includes rolling access to Z.ai's latest flagship models. In September 2026 a developer publishing as ferstar documented that ZCode's checkpoint feature packages the whole workspace, including the complete .git directory, encrypts it and uploads it to Aliyun object storage under a key only Z.ai holds — worth reading before pointing it at a private repository.

What it does

  • Goals: a unit of work for long-running tasks that several agents can work on together
  • Deep GLM-5.3 integration, which Z.ai says is tuned specifically for ZCode's agentic coding loop
  • GLM-5.3-Flash built in for multimodal jobs such as screenshot understanding and image analysis
  • Bot control that lets you drive a running goal from WeChat, Feishu or Telegram
  • Native installers for macOS (Apple Silicon and Intel), Windows (x64 and ARM64) and Linux in beta
  • Three subscription tiers with weekly credit allowances, from 10,000 on Lite to about 140,000 on Max

Getting started

ZCode is a downloadable desktop app, so setup is an installer plus a Z.ai account rather than a package-manager command.

Download the build for your platform

Pick the installer from zcode.z.ai: macOS Apple Silicon or Intel, Windows x64 or ARM64, or a Linux .deb, .rpm or AppImage (Linux support is marked beta).

Sign in and choose a plan

ZCode runs on Z.ai credits. Lite is $12.60 a month for 10,000 credits a week, Pro is $56 for roughly 60,000, and Max is $117.60 for roughly 140,000.

Open a project and create a Goal

Goals are how ZCode tracks long-running work, and several agents can run against one goal. GLM-5.3 drives the coding; GLM-5.3-Flash handles screenshots and other image input.

Decide what leaves the machine

ZCode's checkpoint feature uploads a snapshot of the workspace, including the full .git history, to Z.ai-controlled cloud storage. Review that against your policy before using it on a private or client repository.

Commands and code are distilled from the project's own documentation — always check the official docs for the latest.

When to use it

  • Hand a multi-step refactor or feature to a goal and let several agents work it at once
  • Stay on GLM-5.3 end to end instead of paying per-token for a third-party frontier model
  • Kick off or check on a long build from WeChat, Feishu or Telegram while away from the desk
  • Feed a screenshot of a broken UI straight into the agent via the bundled GLM-5.3-Flash

How ZCode compares

ZCode alongside other open-source autonomous coding agents tools AI/TLDR tracks, ranked by GitHub stars.

ToolStarsWhat it does
opencode★ 208kOpenCode is an open source AI coding agent that runs in your terminal, with built-in build and plan agents and an optional desktop app.
Claude Code★ 146kAnthropic's agentic coding tool for the terminal, IDE and GitHub: it understands your codebase, executes routine tasks, explains code and handles git workflows from natural-language commands.
OpenAI Codex★ 125kOpenAI's cloud and CLI coding agent that writes features, fixes bugs, and proposes code changes across a repo, running tasks in parallel.
Gemini CLI★ 107kAn open-source command-line AI agent from Google that connects your terminal to Gemini models for reading code, editing files, running shell commands, and searching the web.
Pi★ 107kMinimal terminal coding agent harness with four built-in tools, extended through TypeScript extensions, skills and prompt templates instead of forks.
autoresearch★ 96.3kKarpathy's minimal harness that hands a coding agent a single-GPU LLM training script and lets it run fixed five-minute experiments, keeping or discarding each change on its own.
OpenHands★ 88.4kAn open-source AI software-development agent that plans tasks, edits files, runs commands, and tests code, usable from a terminal CLI, a local web GUI, or a Python SDK.
ZCodeZ.ai's desktop coding agent, tuned for GLM-5.3, that runs multi-agent goals and can be driven from a chat app