GLM-5 AI assistant for chat, writing, and programming.
A versatile AI assistant built on GLM-5. Supports chat, writing, coding, and image & document analysis to enhance your learning and productivity.
Released May 25
Freemium
Reviewed by Shawn H. — Founder, AITrustList
Last verified MethodologyAI tools are ranked by traffic signals, not paid placement.

Coding agent for open models that learns your style
A terminal-based AI coding agent designed for open models. It features tool call repairs, 99%+ cache hit, which continuously learns your coding preferences.

Independently written by the AITrustList editorial team.
Other products in the same task, ranked by popularity.
Pricing and commercial details available on the current listing.
What it is, how to use it, and the full listing description.
Command Code is a terminal-based AI coding agent for open models that continuously learns your coding taste. It repairs tool calls on every model, achieves 99%+ cache hit rates, and auto-generates project-level skills from your accepted edits.
Install globally via npm: `npm i -g command-code`. Start a session by running `command-code` in your terminal. Use `/agents` for custom agents, `/memory` for persistent context, and `npx taste push/pull` to share learned skills with your team.
Command Code is a terminal-based AI coding agent built specifically for open models. It validates and repairs tool calls in flight on every model, achieving industry-leading 99%+ prompt cache hit rates. The agent includes taste-1, a continuous reinforcement learning system that learns from every accept, reject, and edit to generate project-level skills automatically.
Key capabilities include:
Interactive CLI, headless -p/--yolo modes, and background sandbox agents
Custom /agents and persistent /memory across sessions
Hackable /skills, /commands, /mcp servers, and plugins
Session sharing and npx taste push/pull for team collaboration
/design command for UI audits, recolor, and reship
Plan review mode with commenting and revision
Support for 50+ models including DeepSeek, Qwen, Kimi, GLM, MiniMax, and Google models
Pricing starts at $1/month for the Go plan. The GOAT plan at $10/month provides up to $100 in credits and access to 50+ models. Provider ($15/mo), Pro ($20/mo), Max 10x ($100/mo), and Max 20x ($200/mo) plans are also available.
Traffic and engagement metrics for commandcode.ai.
Command Code had 1.1M monthly visits in the latest recorded month. It is currently tracked in the Coding category on AITrustList.
| Keyword | Traffic | Cost Per Click |
|---|---|---|
| command code | 74.5K | $ 1.79 |
| commandcode | 41.7K | $ 1.77 |
| command code goat | 5.0K | -- |
| command goat | 2.1K | -- |
| glyph cluster | 7.6K | -- |
0.00 out of 5
Based on 0 reviews
No published reviews yet.
Questions and answers included with this listing.
Command Code is a terminal-based coding agent for open models that continuously learns your coding taste via taste-1. It repairs tool calls in flight, achieves 99%+ cache hit rates, and ships/fixes/tests/refactors with patterns you keep.
Command Code supports models listed in the docs including Anthropic, OpenAI, Google, xAI, DeepSeek, Qwen, Kimi, GLM, MiniMax, and more. New vendors are added regularly.
Every accept, reject, and edit is a signal. Command Code distills those into project-level /skills and personal /memory so future sessions open with your preferred conventions. No rules to write, no prompts to maintain.
A free tier is available for solo developers. Pro and Team plans add seats, more compute, and shared taste registries. See the pricing page for full breakdown.
Official and listing-specific external resources.