Claude Code is Anthropic's agentic coding tool for the terminal, IDEs, a desktop app and the web, billed through a Claude plan or an API account. HagiCode is a desktop workbench that runs Claude Code as its executor and adds parallel sessions, proposals and AI commits around it.
Claude Code is Anthropic's coding agent: it reads your codebase, edits files, runs commands and works with Git, in the terminal, in VS Code and JetBrains, in a desktop app or on the web. HagiCode is a desktop and web workbench that runs the Claude Code as its executor. It does not replace Claude Code, and it does not host or resell Claude models.
Claude Code already covers a lot on its own, including subagents, background agents, hooks, skills and CI use. If you work in one place at a time and are happy with its own surfaces, stay with Claude Code alone. HagiCode is for people who want written proposals before changes, one workspace across repositories, and one view of many sessions and several different Agent CLIs.
| Aspect | Claude Code | HagiCode |
|---|---|---|
| What it is | Anthropic's agentic coding tool; one engine behind the CLI, IDE extensions, desktop app and web | Desktop and web workbench that runs Agent CLIs, Claude Code among them; not a model or an agent itself |
| Interface and platforms | Terminal CLI on macOS, Windows (native or WSL) and Linux; VS Code, JetBrains, desktop app and claude.ai/code | Desktop app for Windows, macOS and Linux; Docker Compose container; Microsoft Store edition |
| Pricing and licensing | Needs a Pro ($20 per month), Max (from $100), Team or Enterprise plan, a Console account, or Bedrock, Vertex or Foundry; free claude.ai plan excluded (as of October 2026) | Desktop and Container editions are free; Store edition is free with an optional one-time TurboEngine purchase |
| Model and provider access | Claude models through your Anthropic plan or a third-party cloud provider | Hosts no models; uses whatever your Claude Code install is signed in to; optional OmniRoute model catalog |
| Parallel work | Subagents, plus background agents to run several sessions and watch them from one screen | Several sessions side by side, each an Agent with a visible status, across any Agent CLI |
| Spec and planning workflow | Plan mode, CLAUDE.md memory, skills and hooks; no proposal archive of its own | OpenSpec proposals: propose, validate, apply, archive; free editions allow 3 concurrent proposals |
| Git and commit handling | Stages changes, writes commit messages, creates branches and opens pull requests when asked | AI-composed commits, repository locked while committing, per-repo targeting with Monospecs |
| Setup effort | One native installer, Homebrew, WinGet or npm command, then browser sign-in | Needs Claude Code installed and signed in first, then HagiCode Desktop on top |
| Best for | Developers who want the agent in their terminal or editor, with CI and scheduled tasks | Spec-driven work, many parallel sessions, multi-repo projects, mixing several Agent CLIs |
The terminal CLI, the VS Code and JetBrains integrations, the desktop app and the web version all use the same Claude Code engine, so CLAUDE.md files, settings and MCP servers carry across them. Remote Control lets you continue a local session from a phone or browser, and claude --teleport pulls a web session into your terminal.
The recommended install is curl -fsSL https://claude.ai/install.sh | bash on macOS, Linux and WSL, or irm https://claude.ai/install.ps1 | iex in PowerShell. Homebrew (brew install --cask claude-code), WinGet (winget install Anthropic.ClaudeCode) and npm (npm install -g @anthropic-ai/claude-code) also work. Run claude --version, then claude and follow the browser sign-in; native installs auto-update.
The free claude.ai plan does not include Claude Code, so you need a paid plan, a Console account or a cloud provider.
CLAUDE.md sets standards for every session, and Claude Code can read an AGENTS.md instead. Skills package repeatable workflows, hooks run shell commands before or after its actions, and auto memory saves what it learns between sessions.
Subagents split a task across agents with a lead coordinating, and background agents run several full sessions you can watch from one screen. claude -p suits pipes and CI, GitHub Actions and GitLab CI/CD are documented, routines and /loop handle scheduled work, and the Agent SDK builds custom agents on the same tools.
claude -p or scheduled routines;HagiCode sits on top of the CLI. It starts your installed Claude Code, so your Anthropic plan, sign-in and usage limits stay with you, and HagiCode adds the session, proposal and repository layers around it.
Nothing is locked in. Your repositories, CLAUDE.md, skills and settings stay where they are, and you can run claude on the same project whenever you like.
claude --version.claude once and sign in, so it is authenticated before HagiCode uses it.claude command on your PATH.claude in the same folder; nothing needs to be undone.Each Claude Code session becomes an Agent with its own context and a status you can scan. That sits alongside Claude Code's own subagents and background agents: HagiCode adds one list for sessions across folders and across different Agent CLIs.
A task can begin as an OpenSpec proposal that you write, validate, apply and archive. Claude Code implements within it, and the proposal keeps the decision record that CLAUDE.md and chat history do not.
HagiCode drafts commit messages from the changes Claude Code made and locks the repository while committing. The embedded Code Server editor opens the touched files beside the conversation.
Monospecs lists the repositories in a multi-repo workspace so one Claude Code task sees all of them. Vault keeps reference code and notes across projects, complementing what auto memory keeps per project.
Claude Code's tool calls appear 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.
When Claude Code needs a decision, it can ask a structured question during the run. HagiCode shows it in the conversation, the session changes from executing to waiting for a reply, and your answer lets the same run continue. If the run has ended, for example after a restart, the answer goes in as a normal message in the same session.
While a run is executing, Claude Code may be waiting on subagents or monitors. HagiCode lists that work with its status, so you can see what the run is waiting for instead of guessing. The list is kept in memory for the current run and is recovered if you refresh the page.
claude directly at any time.curl -fsSL https://claude.ai/install.sh | bash; in Windows PowerShell run irm https://claude.ai/install.ps1 | iex. Homebrew (brew install --cask claude-code), WinGet (winget install Anthropic.ClaudeCode) and npm (npm install -g @anthropic-ai/claude-code) also work. Then run claude --version and claude. Docs are at code.claude.com/docs.claude on the same repositories.HagiCode is an agentic coding workspace. Rich, convenient UI interactions make planning, agent execution, and project work approachable and easy to navigate, creating an engaging development experience.
Explore an engaging development experience through rich, convenient UI interactions.
HagiCode is an agentic coding workspace. OpenSpec workflows plan the change, multi-agent execution carries it out, and Hero Scenario makes the long run visible, so ideas become shipped software.
© 2026 HagiCode | Powered By hagilight-starlight@0.21.0 | Branch: main | Commit: 951128b | Built:
