CodeBuddy vs HagiCode: Features and Comparison

HagiCode is a proposal-driven AI coding workbench that turns a request into a reviewable proposal, runs it through agent CLIs such as Claude Code, Codex and GitHub Copilot, and keeps the outcome as team knowledge.

CodeBuddy Code is Tencent Cloud's terminal coding agent, installed from npm, Homebrew or a native binary. HagiCode is a desktop workbench that runs CodeBuddy as its executor and adds parallel sessions, proposals and AI commits around it.

Install HagiCode Desktop

English2026-10-16

Short answer

CodeBuddy Code is the command-line coding agent from Tencent Cloud's CodeBuddy product line, which also includes an IDE and editor plugins. HagiCode is a desktop and web workbench that runs CodeBuddy as one of its executors. It does not replace CodeBuddy and does not host models.

If you want an agent in your terminal, you use your CodeBuddy account, and you work on one task at a time, CodeBuddy alone is enough. HagiCode is for the point where you want several CodeBuddy tasks running side by side, written proposals before code changes, or one workspace across several repositories.

CodeBuddy vs HagiCode at a glance

AspectCodeBuddy CodeHagiCode
What it is

Terminal AI coding agent from Tencent Cloud; sibling products are the CodeBuddy IDE and editor plugins

Desktop and web workbench that runs Agent CLIs, CodeBuddy among them; not a model or an agent itself

Interface and platforms

Terminal, with -p for scripted runs; native binary for macOS, Linux and Windows (beta), or npm and Homebrew

Desktop app for Windows, macOS and Linux; Docker Compose container; Microsoft Store edition

Pricing and licensing

Free, Pro (about $10 per month) and Team (about $40 per seat) plans, credit-based; as of October 2026, per Tencent Cloud's billing page

Desktop and Container editions are free; Store edition is free with an optional one-time TurboEngine purchase

Model and provider access

Models come with your CodeBuddy account and site; extra models via models.json with an OpenAI-compatible endpoint

Hosts no models; uses whatever your CodeBuddy install is signed in to; optional OmniRoute model catalog

Parallel work

Sub-agents inside a session, and Agent Teams exposed over ACP

Several sessions side by side, each an Agent with a visible status

Spec and planning workflow

Plan permission mode, /init for project context, hooks and skills

OpenSpec proposals: propose, validate, apply, archive; free editions allow 3 concurrent proposals

Git and commit handling

Built-in Git operations; can generate commit messages in the terminal

AI-composed commits, repository locked while committing, per-repo targeting with Monospecs

Setup effort

One npm, Homebrew or script install, then sign in to the China, International or enterprise site

Needs CodeBuddy installed and signed in first, then HagiCode Desktop on top

Best for

Terminal users in Tencent Cloud's ecosystem, scripts and CI

Parallel tasks, proposal-driven work and multi-repo projects

What CodeBuddy does well

Several install routes

CodeBuddy Code installs with npm install -g @tencent-ai/codebuddy-code (Node.js 18.20 or later), or through Homebrew with brew install Tencent-CodeBuddy/tap/codebuddy-code. Homebrew does not need Node.js, and a beta native binary does not either, and codebuddy install migrates an npm install to it.

The command is codebuddy (the docs also mention a cbc alias); check it with codebuddy --version. codebuddy update upgrades it, and auto-update is on by default.

Sign-in by region and by organisation

On first launch you choose where to sign in: the China site, the International site, an enterprise domain for dedicated deployments, or iOA for Tencent staff. The China and International sites expose different model sets, so check which one your account belongs to before comparing models.

Scriptable and permission-aware

codebuddy -p runs a single prompt and accepts piped input, so it works in shell pipelines, for example cat error.log | codebuddy -p "...". Shift+Tab cycles permission modes (default, bypass, accept edits, plan), and -y skips confirmations for runs that need file or command access.

Extensibility and enterprise reach

Docs cover MCP servers, sub-agents, plugins, skills, hooks, Python and TypeScript SDKs, GitHub Actions and a web UI. Models can be added in ~/.codebuddy/models.json or per project, and enterprise deployments can run in containers. codebuddy --acp exposes it as an ACP server with a context-window option and Agent Teams status, which is the interface HagiCode uses.

When to choose CodeBuddy, and when to choose HagiCode

Stay with CodeBuddy when

  • you work in a terminal and one task at a time is enough;
  • you are already in Tencent Cloud's ecosystem or on an enterprise CodeBuddy domain;
  • you want scripted runs with codebuddy -p, piped input or GitHub Actions;
  • you want the lightest route: a Homebrew or native-binary install with no extra application;
  • you prefer its own permission modes, hooks and plugins to a separate workbench.

Reach for HagiCode when

  • you run several CodeBuddy tasks at once and want them in one view;
  • you want a reviewed proposal before the agent edits code, archived afterwards;
  • a CodeBuddy task needs edits in more than one repository and you want to plan it once;
  • you want CodeBuddy's changes committed with generated messages, without leaving the window for an editor;
  • you use CodeBuddy alongside other Agent CLIs and want one workspace.

Using CodeBuddy together with HagiCode

HagiCode sits on top of the CLI. It starts your CodeBuddy install over ACP, so your account, region and credits stay with CodeBuddy and HagiCode does not take over its sign-in.

Nothing is locked in. Your repositories and your .codebuddy settings stay where they are, and you can run codebuddy on the same project whenever you like.

  • Install CodeBuddy Code with npm, Homebrew or the native installer, and confirm it with codebuddy --version.
  • Run codebuddy once and sign in on the site that matches your account (China, International or enterprise).
  • Install HagiCode Desktop for Windows, macOS or Linux, or use the Docker Compose or Microsoft Store edition.
  • In HagiCode's Agent CLI settings, pick CodeBuddy as the executor; HagiCode looks for the codebuddy command on your PATH.
  • Open a repository and start a session or an OpenSpec proposal; CodeBuddy's tool calls appear in HagiCode's unified view.
  • To go back, close HagiCode and run codebuddy in the same folder; nothing needs to be undone.

What HagiCode adds on top of any agent

Parallel sessions and Agents management

Each CodeBuddy session becomes an Agent with its own context and a status you can scan, instead of one more terminal running codebuddy. That helps most when you start tasks across several folders.

OpenSpec proposals

A task can begin as an OpenSpec proposal you write, validate, apply and archive. CodeBuddy implements within it, and the proposal keeps the reasoning that a terminal scrollback loses.

AI commit and Code Server

HagiCode drafts the commit message from CodeBuddy's changes and locks the repository while it commits. The embedded Code Server editor opens the files CodeBuddy touched next to the conversation.

Monospecs and Vault

Monospecs lists the repositories of a multi-repo workspace so a CodeBuddy task sees all of them. Vault keeps reference projects and notes available to later sessions, including those started in a fresh codebuddy process.

Model routing, tool-call view and usage

CodeBuddy's tool calls are shown in the same typed shape as every other CLI, live and in history. OmniRoute can supply a model catalog per Agent CLI, and an optional TokenMonitor readout in the status bar shows today's tokens and cost.

The bottom line

  • CodeBuddy Code is the agent in your terminal; HagiCode is the workbench that runs it.
  • Use CodeBuddy alone for solo, single-task or scripted work; add HagiCode for parallel tasks, proposals and multi-repo changes.
  • CodeBuddy must be installed and signed in first, and you can go back to running codebuddy directly at any time.

Frequently asked questions

How do I install CodeBuddy Code?

Run `npm install -g @tencent-ai/codebuddy-code` (Node.js 18.20 or later), or on macOS and Linux `brew install Tencent-CodeBuddy/tap/codebuddy-code`. A beta native binary needs no Node.js. Then run `codebuddy --version` and `codebuddy` to sign in. See the docs at codebuddy.ai/docs/cli/installation.

Is CodeBuddy free?

CodeBuddy has a free plan with a small monthly credit allowance, plus Pro and Team plans (Pro about $10 per month, Team about $40 per seat, as of October 2026). Tencent Cloud's billing page does not say which plans include the CLI, so check it before relying on the free tier.

Can I use CodeBuddy with HagiCode?

Yes. Install and sign in to CodeBuddy Code, then choose it as the Agent CLI in HagiCode. HagiCode drives it over ACP, and your account and credits stay with CodeBuddy.

Does HagiCode replace CodeBuddy?

No. HagiCode runs CodeBuddy as its executor and adds sessions, proposals, commits and multi-repo tools around it. You can still run `codebuddy` on the same repositories.

What is the CodeBuddy command?

The command is `codebuddy`, with `cbc` mentioned as a short alias in the docs. The npm package is @tencent-ai/codebuddy-code.

CodeBuddy vs HagiCode: which should I pick?

Pick CodeBuddy alone for terminal, solo or scripted work. Add HagiCode when you want parallel sessions, OpenSpec proposals or multi-repo workspaces on top of CodeBuddy.

Desktop

Localized AI code assistant, privacy protection, enhanced efficiency