Murage with Claude Code

Keep Claude Code and give it a team. Murage runs the real Claude Code on your own Claude plan, under a Chief of Staff who builds a team and puts Codex, Hermes or other engines beside it. Murage is free.

Murage runs this engine as a bot on your team. It’s a separate product, not a partnership.

The short answer

Keep Claude Code for the building. Murage gives it a team for everything else, on the same Claude plan.

Which should you choose?

Murage runs the team

Murage

The app is free
On your computer
The AI you choose

Who does the work: A Chief of Staff and a team

Claude Code runs this bot

Claude Code

Pro $20 a month
Terminal, IDE, desktop, web and mobile
Claude models only

Who does the work: Subagents and agent teams

Download Murage

Checked against official pages and the sources below on September 25, 2026.

Murage sits above Claude Code.

You talk to one Chief of Staff. It builds the team, and each bot runs the engine that suits its job, on the plan or key you already have. An engine is the AI that does a bot’s thinking. To run teams and use tools, pick a tool-capable one: Claude Code, Codex, Hermes, or the built-in Fuigo engine with a supported model. A plain OpenAI-compatible endpoint works for chat and drafting only. Portraits made with Murage’s own avatar generator. See every engine

Claude Code runs inside Murage, as a bot’s engine

Keep Claude Code. Murage makes it a team of specialists.

You already pay for one of the best coding agents. Murage makes it one member of a standing team. Tell your Chief of Staff what you need, and it sends the building to Claude Code bots and the research, writing and follow-up to bots on whichever engine (the AI that does a bot’s thinking) suits. If Claude Code isn’t installed yet, Murage opens a one-line install for you. Then you sign in to the unmodified Claude Code through Anthropic’s own flow, and Murage never touches the login.

What Murage adds on top

A Chief of Staff that takes the brief, plans the job and hands out the work
Teams of named bots, each with its own face, voice, personality and memory on your computer
1,500+ apps (Gmail, Slack, Notion, GitHub) through Flux Router credits or your own connected-apps key
Voice calls with any bot, plus your Chief on Telegram, Slack and Discord: send a voice message and get a voice reply
New bots start on Ask, and on Full access a stop line holds back the few outside actions that wait for your yes, unless you allow it for the task or turn it off. On Hermes and other ACP engines, Ask and the stop line cover only shell commands and file edits
Swap engines whenever you like. Your team, memory and approvals stay put

Why the app is free Every engine Murage runs

Claude Code already has subagents and agent teams. Why add Murage?

They’re excellent, and they’re built for work inside a Claude Code session. Claude hands tasks to subagents on its own, and you can save subagent definitions in files, each with its own model and a memory folder that “survives across conversations” (Anthropic docs). Agent teams go further: several Claude Code sessions that message each other. They’re “experimental and disabled by default”, with “one team per session”, and teammates “cannot spawn their own teammates” (Anthropic docs).

Murage works one level up. Its team lives outside any one session: named bots with faces, voices and memory that are still there tomorrow, on routines, reachable from Telegram, Slack or Discord, and not all on Claude. A Claude Code bot can sit next to a Codex bot and a Hermes bot.

Claude Code subagents Claude Code agent teams

At a glance.

Ten things buyers weigh up, including the ones where they’re ahead.

What matters

Murage

Claude Code

The app itself is free
Before your AI plan or key, which you pay for either way
Yes
Free, with no paid tier
Source
No
Needs a paid Claude plan, from Pro at $20 a month
Source
You choose which AI company does the work
Per bot or per task, from different vendors
Yes
Per bot: Claude Code, Codex, Hermes, a local model and more
No
Anthropic’s Claude models
Source
Named agents that keep their memory
A standing team that lasts beyond one task
Yes
Named bots, each with its own notebook and managed memory
Partly
Saved subagents keep their own memory; agent teams are experimental, one per session
Source
Asks before acting outside, by default
Messaging someone new, posting, paying, deleting
Yes
New bots start on Ask; on Full access the stop line judges what each engine asks
Source
Partly
Permission prompts in sessions; cloud Routines run without stopping for approval
Source
Memory stored on your computer
Not in a vendor’s cloud
Yes
Memory, files and chats on your computer
Yes
CLAUDE.md files, auto memory and subagent memory folders
Source
Open source
A license you can read
Yes
AGPL-3.0
Source
No
Its license reads “© Anthropic PBC. All rights reserved.”
Source
Reach it from your phone
Briefs and approvals away from the desk
Yes
The web app, plus your Chief in Telegram, Slack or Discord
Yes
Remote Control from your phone, and the Claude mobile app
Keeps working with your computer off, today
Hosted or on a server
No
Murage Cloud isn’t open yet
Yes
Routines (at most hourly, with a daily cap) and web sessions on Anthropic’s cloud
Source
Native phone app
An iOS or Android app, not a web app or chat bot
Coming
iPhone and Android app coming
Yes
The Claude mobile app
SOC 2 report
What a procurement team will ask for
No
No SOC 2 report
Yes
A SOC 2 report, through Anthropic’s Trust Center
Source

Each of Claude Code’s marks links the page it comes from, checked September 25, 2026. “Not stated” means we couldn’t find it on the page named, not that it’s missing.

What a year costs.

The Murage app is

Free

Your Claude Pro plan costs $240 a year either way. Murage adds nothing to that bill, and gives you a Chief of Staff and a team.

A year, one person, entry plan

Claude Code on its own · Claude Pro, the cheapest plan that includes it$240

Claude Pro with Murage · same plan, plus a team$240 + free app

Already on Claude? Then Claude Code is included, and Murage runs it on that same plan. Murage runs it the way you’ve set it up, so that cost stays the same. Prices from official pricing pages, billed monthly, checked September 25, 2026.

Where your work lives.

Murage

Your computer

MemoryFilesChats

Your memory, files and chats are stored on your computer. What a bot sends to your AI, and to any app or service you connect, goes to them.

Claude Code

Your computer

MemoryFiles

Optional cloud

It works in your folders on your computer. Web sessions and Routines run on Anthropic’s cloud.

Who does the work.

You brief one Chief of Staff. It builds a team of specialists for the job, and the team keeps its memory for next time. You can put a reviewer on any team to check work before it reaches you. Portraits made with Murage’s own avatar generator.

Claude Code

One session hands work to subagents; experimental agent teams let sessions message each other.

What Murage adds for a Claude Code user.

The brief

“Go through my inbox before I get in. Draft replies in my voice. Send nothing.”

What came back

Twelve emails came in. The team drafted eight replies, set four aside, and left two questions only the owner could answer. Nothing was sent.

Needs from you, in the team’s own report

Dana Brooks, Acme Studio · RenewalThe prior contract terms
Jordan Lee, Brightline Design · Pricing for 20 seatsThe per-seat price, and whether we offer a pilot
Priya S. · Export to PDF?Confirm the exact steps
Ben C. · Refund requestMoney. Check the billing record first

Replay of a real run on demo data. The company is made up. Engine: Flux Auto, worked 1m 41s. Watch the run

It runs the real Claude Code

You sign in to the unmodified Claude Code with your own Claude subscription, through Anthropic’s own sign-in. Murage gives it a seat on a team and never handles your login.

A Chief of Staff above it

Brief one Chief of Staff in plain English. It builds the team and decides which bot does what, so you are not steering sessions one by one.

Every engine on one team

Put Claude Code on the bots that build, and Codex, Hermes, Grok Build or an API key on the rest. Claude Code on its own runs Claude models. Swap engines later and nothing else changes.

Built for the rest of the business

Claude Code is built around code and repositories. Your team also has to reply to leads, write the brief and chase the invoice. Pick from 68 ready-made teams, connect 1,500+ apps, call any bot and talk, and approve from Telegram.

Where Claude Code is ahead.

Fair’s fair. If one of these matters most to you, it may be the better choice today.

1
The deepest coding agent for Claude
Built by the company that trains the model, with subagents, agent teams, IDE integrations and dynamic workflows.
2
Runs in the cloud
Routines and web sessions keep going on Anthropic’s cloud with your laptop closed. Murage runs while your computer is on, until Murage Cloud opens.
3
Everywhere developers work
Terminal, IDEs, desktop, web and mobile, plus Remote Control from your phone.
4
A huge community
Its repository has about 148,000 GitHub stars (checked September 27, 2026), and MCP connectors reach the tools developers use.

Which should you choose?

Keep Claude Code for the building. Murage gives it a team for everything else, on the same Claude plan.

Choose Murage if

You want Claude Code on a standing team with Codex, Hermes or local models, run by a Chief of Staff you brief once
You want to reach it from Telegram, Slack or Discord, and have new bots ask before they message, post, pay or delete
You want Claude Code on a team that also handles sales, ops and research
You want everyone on your staff to have it, with no Murage seat fees

Download Murage

Choose Claude Code if

You want the agent inside your terminal, IDE and the Claude app, with nothing in between
You need cloud Routines that keep running with your laptop closed, today
You want everything from the company that makes the model

Still deciding? Read how we check that a job is finished, then try one of your own.

Every detail.

Question	Murage	Claude Code
Price to start	Free. No paid tier	A paid Claude plan: Pro $20 a month ($17 billed yearly). Not on Free
Usage	Nothing from us. Your AI provider bills as usual. Flux Router, Ferrox Labs’ optional credit-billed service, covers AI models, connected apps, voice, images and web search if you use it; your own keys work instead	Your plan’s limits; a team of bots uses them faster. Enterprise pays usage at API rates
Where it runs	On your computer	Your terminal, IDE or the Claude desktop app, plus web sessions and Routines on Anthropic’s cloud
Which AI	The engine you choose, per bot: Claude Code, Codex, Hermes, Grok Build, Cursor, Google Antigravity, Kimi, Qwen, Droid, OpenCode, Pi, Fuigo (built in), local models (Ollama, LM Studio) and any OpenAI-compatible API (chat only). OpenClaw is coming.	Anthropic’s Claude models
Memory	On your computer, open to read and edit	CLAUDE.md files plus auto memory
Open source	Yes, AGPL-3.0	No. Its license reads “© Anthropic PBC. All rights reserved.”
Keeps working with your laptop closed	Not today. It runs while your computer is on; Murage Cloud isn’t open yet	Yes, for Routines (at most hourly, with a daily cap) and web sessions
Approvals	New bots start on Ask. On Full access, a bot stops before messaging anyone new, posting publicly, paying, or deleting outside its folder, unless you have allowed it for that task. Replies in an existing conversation go ahead on Auto and Full access. No limits turns the stop line off. On Claude Code and Codex, the stop line also covers tools from MCP servers you add. On Hermes and other ACP engines, Ask and the stop line cover only what the engine asks Murage about: Hermes asks before shell commands and file edits, but not before MCP or connected-app tools.	Permission rules plus modes such as ask and auto. Routines run without asking
Teams	A Chief of Staff builds and runs a team for each job	Subagents Claude hands work to on its own, saved subagents with their own memory, and experimental agent teams (off by default, one per session)
Voice	Call any bot in the app and just talk; bots leave voice notes. Send your Chief a voice message in Telegram, Slack or Discord and get a voice reply	Voice dictation in the CLI; no voice calls

Questions about Murage and Claude Code.

Isn’t Murage just a wrapper around Claude Code?

Partly, and on purpose. Murage runs the real Claude Code, because Anthropic builds the best agent for Claude. What Murage adds is what one Claude Code session doesn’t have: a standing team of named bots that remember the job tomorrow, Codex and Hermes on the same team, a stop line you can read in the source, your Chief of Staff on Telegram, Slack and Discord, and voice calls. If you only write code in a terminal, Claude Code on its own may be all you need.

Is this allowed under Anthropic’s terms?

Yes. Murage never signs you in, and never sees, stores or passes on your Claude login. It helps you install the official Claude Code, you sign in to Claude Code yourself on Anthropic’s own screen, and Murage runs it as a bot’s engine. Anthropic’s policy says its rules don’t “prevent an end user from signing in to the unmodified Claude Code binary with their own Claude subscription” (Claude Code legal and compliance, checked September 26, 2026).

Your usage counts against your own plan, as it would if you ran Claude Code yourself. A team of bots gets through that allowance faster than one chat, so for heavy work you can give any bot an Anthropic API key instead. Murage is not affiliated with or endorsed by Anthropic.

Do I need another Claude plan for Murage?

No. You sign in to the unmodified Claude Code with your own Claude subscription. If Claude Code isn’t installed, Murage opens a one-line install for you to paste first. Murage adds no fee on top, but a team of bots uses your plan’s allowance faster than one chat.

Does Murage replace Claude Code?

No, it runs it. Claude Code becomes the engine under one or more of your bots, and Murage adds the Chief of Staff, the team, memory and approvals on top.

Can I switch engines later?

Yes. A bot’s engine is a setting you can change at any time. Its memory, skills, routines and approvals live in Murage, not in the engine, so they stay put when you switch.

Can my staff use it too?

Yes. Each person runs Murage on their own computer, with their own AI sign-in. Build a team once, export it, and everyone imports the same bots, skills and routines; logins never travel. Bots can post drafts to your team’s Slack channels through connected apps, with your approval. Shared always-on workspaces will come with Murage Cloud, which isn’t open yet.

Is Murage made by Anthropic?

No. Murage is an independent open-source project. It is not affiliated with or endorsed by Anthropic.

Sources, checked September 25, 2026

Murage is made by Ferrox Labs, a small team. Spotted something out of date? Tell us and we’ll fix it.

Claude and Claude Code are trademarks of their respective owners. Murage is not affiliated with or endorsed by Anthropic; the names and logos are used only to identify the products compared.

Other comparisons.

Keep Claude Code. Add the team.

Give your first job to Murage.

Download the free app, connect the AI you already pay for, and tell your Chief of Staff what needs doing. Plan on about ten minutes from install to a working team.

Free · Download Murage

No account needed · Runs on the AI plan you already pay for

Murage with Claude Code

Murage runs this engine as a bot on your team. It’s a separate product, not a partnership.

The short answer

Keep Claude Code for the building. Murage gives it a team for everything else, on the same Claude plan.

Which should you choose?

Murage runs the team

Murage

The app is free
On your computer
The AI you choose

Who does the work: A Chief of Staff and a team

Claude Code runs this bot

Claude Code

Pro $20 a month
Terminal, IDE, desktop, web and mobile
Claude models only

Who does the work: Subagents and agent teams

Download Murage

Checked against official pages and the sources below on September 25, 2026.

Murage

Claude Code

Price to start

Free. No paid tier

A paid Claude plan: Pro $20 a month ($17 billed yearly). Not on Free

Usage

Nothing from us. Your AI provider bills as usual. Flux Router, Ferrox Labs’ optional credit-billed service, covers AI models, connected apps, voice, images and web search if you use it; your own keys work instead

Your plan’s limits; a team of bots uses them faster. Enterprise pays usage at API rates

Where it runs

On your computer

Your terminal, IDE or the Claude desktop app, plus web sessions and Routines on Anthropic’s cloud

Which AI

The engine you choose, per bot: Claude Code, Codex, Hermes, Grok Build, Cursor, Google Antigravity, Kimi, Qwen, Droid, OpenCode, Pi, Fuigo (built in), local models (Ollama, LM Studio) and any OpenAI-compatible API (chat only). OpenClaw is coming.

Anthropic’s Claude models

Memory

On your computer, open to read and edit

CLAUDE.md files plus auto memory

Open source

Yes, AGPL-3.0

Keeps working with your laptop closed

Not today. It runs while your computer is on; Murage Cloud isn’t open yet

Yes, for Routines (at most hourly, with a daily cap) and web sessions

Approvals

New bots start on Ask. On Full access, a bot stops before messaging anyone new, posting publicly, paying, or deleting outside its folder, unless you have allowed it for that task. Replies in an existing conversation go ahead on Auto and Full access. No limits turns the stop line off. On Claude Code and Codex, the stop line also covers tools from MCP servers you add. On Hermes and other ACP engines, Ask and the stop line cover only what the engine asks Murage about: Hermes asks before shell commands and file edits, but not before MCP or connected-app tools.

Permission rules plus modes such as ask and auto. Routines run without asking

Teams

A Chief of Staff builds and runs a team for each job

Subagents Claude hands work to on its own, saved subagents with their own memory, and experimental agent teams (off by default, one per session)

Voice

Call any bot in the app and just talk; bots leave voice notes. Send your Chief a voice message in Telegram, Slack or Discord and get a voice reply

Voice dictation in the CLI; no voice calls