ZCode 3.0: GLM-5.2 optimized, better multi-agent collaboration

Simple, Fast, Vibe‑Ready.

ZCode combines the best AI agents with your existing tools so you can plan, code, review, and deploy without friction. The official harness for GLM‑5.2.

Download ZCode .dmg (Apple Silicon) View all downloads

v3.7.5macOS Windows Linux Beta

Create an intelligent Gomoku (Five-in-a-Row) game… gomoku-ai feat/gomoku-ai

Working

Ask for follow-up changes

GLM-5.2 Max

Git tools

Changes +0 −0

feat/gomoku-ai

Goal In progress

Gomoku vs. AI — implement computer moves with a heuristic algorithm

0/5 · 0:00 · 0 tokens

Initialize board, piece rendering, and the 15×15 grid layout
Implement player move interaction and win-detection logic
Integrate a heuristic AI for automatic computer moves
Adapt board scaling and portrait/landscape layout for mobile
Add rules copy, restart entry points, and empty-state guidance

Capabilities

Stay on the Frontier

From understanding legacy monoliths to shipping real-time features, ZCode keeps every engineer on the frontier of software development.

Long-running tasks

Use Goals to manage complex work with continuous planning, execution, and verification. ZCode decomposes the objective, keeps the plan visible, and checks its own work — step by step, token by token.

Continuous planning, execution, and verification
Live progress, elapsed time, and token budget
Resume long sessions without losing state

Bot control

Start and steer ZCode from WeChat, Feishu, or Telegram so work can keep moving anywhere. Kick off a task from your phone, review the diff, and send the follow-up — the agent keeps running at your desk.

Launch tasks from chat, wherever you are
Steer running agents with plain messages
Progress and results delivered back to the thread

Deep GLM-5.2 integration

Optimized for GLM-5.2 across reasoning, coding, and multi-agent collaboration. The model and the harness are tuned together, so agentic coding runs faster and steadier.

Tuned reasoning for long-horizon coding work
Multi-agent collaboration out of the box
Rolling access to the latest flagship models

ZCode Goal panel tracking progress on a streaming Markdown renderer task

ZCode session being steered from WeChat, Feishu, and Telegram bots

GLM-5.2 ranking on a code arena leaderboard inside ZCode

Z.AI

Code with GLM Coding Plan

GLM is tuned for ZCode, making agentic coding faster and steadier.

GLM Coding

Lite

Built for lightweight iteration on small repo

$12.6/ month ~~$18~~ −30%

10,000 Credits / week1×

Rolling access to the latest flagship models and features
Supports 20+ agent tools, including ZCode, Claude Code, and more
Default data privacy

Choose Lite

Popular

GLM Coding

Pro

Built for day-to-day development on mid-sized repo

$56/ month ~~$80~~ −30%

6× Lite usage6×

All Lite benefits
Priority access to the latest flagship models and features
Includes a curated selection of MCP tools
Faster generation speeds

Choose Pro

Max Usage

GLM Coding

Max

Built for advanced users working on mid-to-large repo

$117.6/ month ~~$168~~ −30%

14× Lite usage14×

All Pro benefits
First access to the latest flagship models and features
Dedicated resources during peak times

Choose Max

Prices and plan benefits may change. Final details are shown on z.ai.

All Downloads

Every platform. One harness.

All ZCode installers for every platform — currently shipping v3.7.5.

macOS

macOS (Apple Silicon) .dmg v3.7.5 macOS (Intel) .dmg v3.7.5

Windows

Windows (64-bit) .exe v3.7.5 Windows (ARM64) .exe v3.7.5

Linux

Beta

Linux x64 .deb v3.7.5 Linux x64 .AppImage v3.7.5 Linux ARM64 .deb v3.7.5 Linux ARM64 .AppImage v3.7.5

Ship at the speed of thought.

Download ZCode and put GLM-5.2 to work on your repo tonight.

Download ZCode.dmg (Apple Silicon) Read the changelog