Skip to content
Back to search

moi

πŸ’Ύ MOI: agent-agnostic generative UI workspace. Give your agent a skill and let it build software around itself.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/moi-computer/moi with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/moi-computer/moi'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/molefrog--moi/badge.svg)](/listings/molefrog--moi)

HTML

<a href="/listings/molefrog--moi"><img src="/api/listings/molefrog--moi/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Make sure [Bun](https://bun.sh) 1.3 or newer is installed. moi uses it to run the web server and bundle the dynamic UI.

Install the `moi-computer` package from npm and start the web UI:

bun i -g moi-computer
moi start        # http://localhost:13337

Then connect your agent: [Claude Code and Codex](#claude-code-and-codex), [OpenClaw](#openclaw), or [Hermes](#hermes).

Learn how to:

- [Run moi as a service](#run-moi-as-a-service) - [Update moi](#update-moi)

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md Β· README Β· 228 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. ```
  2. β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ β–ˆβ–ˆ
  3. β–ˆβ–ˆβ–‘β–‘β–ˆβ–ˆβ–‘β–‘β–ˆβ–ˆ β–ˆβ–ˆβ–‘β–‘β–ˆβ–ˆ β–ˆβ–ˆ
  4. β–ˆβ–ˆβ–‘β–‘β–ˆβ–ˆβ–‘β–‘β–ˆβ–ˆ β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ β–ˆβ–ˆ
  5. THE UI FOR YOUR AI
  6. ```
  7. [![npm version](https://img.shields.io/npm/v/moi-computer?label=npm&labelColor=555&color=E85D2A)](https://www.npmjs.com/package/moi-computer)&nbsp;&nbsp;[![Discord](https://img.shields.io/badge/Discord-Join%20server-5865F2?logo=discord&logoColor=white&labelColor=555)](https://discord.gg/9kj3y9Pfk)
  8. An extendable visual workspace where your AI agents build their own UI.
  9. See [moi.computer](https://moi.computer) for what it is and how it works.
  10. Moi gives your agent of choice a UI it can reshape on the fly and grow into your
  11. personal software. Under the hood, the agent writes and embeds live components
  12. wired to real data: APIs, local files and commands, MCP servers.
  13. Works with the harness you already use:
  14. <p align="center">
  15. <a href="#claude-code-and-codex"><img src="assets/harnesses/claude.svg" width="16" alt="" /> <strong>Claude Code</strong></a>&nbsp;&nbsp;&nbsp;&nbsp;
  16. <a href="#claude-code-and-codex"><img src="assets/harnesses/codex.png" width="16" alt="" /> <strong>Codex</strong></a>&nbsp;&nbsp;&nbsp;&nbsp;
  17. <a href="#openclaw"><img src="assets/harnesses/openclaw.svg" width="16" alt="" /> <strong>OpenClaw</strong></a>&nbsp;&nbsp;&nbsp;&nbsp;
  18. <a href="#hermes"><img src="assets/harnesses/hermes.png" width="16" alt="" /> <strong>Hermes</strong></a>
  19. </p>
  20. # Features
  21. - **Bring your own agent**. Connect the agent you already use.
  22. - **Workspace**. The folder where your agent (e.g. Claude Code) runs and keeps
  23. its data. It gets a special skill that teaches it how to communicate with Moi.
  24. - **Theme**. The agent can modify the appearance of the workspace for you:
  25. fonts, color scheme and more.
  26. - **Scratchpad**. A shared canvas where you ideate together with the agent: ask
  27. it to draw diagrams, add feedback or comments, visualize knowledge and plans.
  28. - **Overview**. A dynamic home made of small widgets wired to the data from
  29. your workspace. They can be rearranged, modified, and completely customized.
  30. - **Views**. An app embedded in the workspace that lives in its own tab. Build
  31. complex interfaces: CRMs, task trackers, second brain storage, customer
  32. support views, etc.
  33. <table>
  34. <tr>
  35. <td align="center"><img src="assets/screenshots/moi-moi-dash-2.webp" alt="Overview" /><br /><sub>Widgets wired to live project stats</sub></td>
  36. <td align="center"><img src="assets/screenshots/moi-magazin-scratch-2.webp" alt="Scratchpad" /><br /><sub>Order funnel on the scratchpad, live from Postgres</sub></td>
  37. <td align="center"><img src="assets/screenshots/moi-magazin-products-2.webp" alt="Products view" /><br /><sub>Use the agent as a copilot when working with products</sub></td>
  38. </tr>
  39. <tr>
  40. <td align="center"><img src="assets/screenshots/moi-magazin-dash-2.webp" alt="Store dashboard" /><br /><sub>E-commerce store dashboard connected to Postgres and Resend</sub></td>
  41. <td align="center"><img src="assets/screenshots/moi-view-screen.webp" alt="Roadmap view" /><br /><sub>Build a Roadmap view with epics and pull requests</sub></td>
  42. <td align="center"><img src="assets/screenshots/moi-scratch.webp" alt="Scratchpad canvas" /><br /><sub>Sketching architecture together with the agent</sub></td>
  43. </tr>
  44. <tr>
  45. <td align="center"><img src="assets/screenshots/moi-color-grading.webp" alt="Color grading review" /><br /><sub>Custom view for color grading</sub></td>
  46. <td align="center"><img src="assets/screenshots/moi-home.webp" alt="Home screen" /><br /><sub>Home screen with your workspaces</sub></td>
  47. <td align="center"><img src="assets/screenshots/moi-linear.webp" alt="Issues view" /><br /><sub>Issues view, a lightweight Linear-like tracker</sub></td>
  48. </tr>
  49. </table>
  50. # Quick start
  51. Make sure [Bun](https://bun.sh) 1.3 or newer is installed. Moi uses it to run the web server and bundle the dynamic UI.
  52. Install the `moi-computer` package from npm and start the web UI:
  53. ```sh
  54. bun i -g moi-computer
  55. moi start # http://localhost:13337
  56. ```
  57. Then connect your agent: [Claude Code and Codex](#claude-code-and-codex),
  58. [OpenClaw](#openclaw), or [Hermes](#hermes).
  59. Learn how to:
  60. - [Run Moi as a service](#run-moi-as-a-service)
  61. - [Update Moi](#update-moi)
  62. # How it works
  63. Moi runs locally as a CLI and web UI connected to your agent. The workspace
  64. skill teaches the agent how to call the CLI, which updates the UI using your
  65. files, commands, APIs, and MCP servers.
  66. The core is `moi` command that both agent and you can use to modify the workspace.
  67. Everything gets propagated to the UI instantly.
  68. ```mermaid
  69. %%{init: {"theme": "base", "themeVariables": {"primaryColor": "#ffffff", "primaryTextColor": "#111111", "primaryBorderColor": "#111111", "lineColor": "#737373", "edgeLabelBackground": "#ffffff"}}}%%
  70. flowchart LR
  71. Agent["Your agent, e.g. Claude Code"] -->|"uses moi skill"| CLI["moi CLI"]
  72. CLI -->|"updates"| UI["UI"]
  73. ```
  74. Moi also stores the all widgets and views as code to keep the workspace filesystem
  75. as a source of truth.
  76. Updates installed from the UI restart Moi and reload the browser automatically.
  77. The app checks npm at most once an hour while it is open, retrying in five
  78. minutes if the registry could not be reached; a closed app makes no network
  79. calls of its own.
  80. # Connect an agent
  81. A workspace is the folder where your agent runs and stores its data. Moi adds
  82. a `.moi/` folder and the skills the agent needs to work with the UI.
  83. You can connect a workspace in either of these ways:
  84. - Open [http://localhost:13337](http://localhost:13337) and create a workspace
  85. or import one that Moi found on your computer.
  86. - Run `moi init` in a folder. Moi detects the harness from the folder's Hermes
  87. profile, OpenClaw agent, or Claude Code or Codex history. If detection is
  88. ambiguous, it asks you to choose a harness.
  89. ## Claude Code and Codex
  90. Open Claude Code or Codex in your project folder and paste this prompt:
  91. ```
  92. Set up the Moi workspace for this project. Fetch https://moi.computer/INSTALL.md, and follow the steps.
  93. ```
  94. To set it up manually, run:
  95. ```sh
  96. moi init --harness=claude-code # or --harness=codex
  97. moi start # or run Moi as a service
  98. ```
  99. The project folder becomes the workspace. You can create as many workspaces as
  100. you need.
  101. ## OpenClaw
  102. [OpenClaw](https://openclaw.ai/) is a self-hosted harness for running one or
  103. more always-on agents. Its channels connect agents to Slack and other
  104. messaging services.
  105. Moi connects each OpenClaw agent to one workspace. See the
  106. [OpenClaw documentation](https://docs.openclaw.ai/cli/agents) to add or manage
  107. agents.
  108. Run `moi openclaw init` to connect your agent. If Moi finds more than one, it
  109. lists them so you can rerun the command with the agent you want:
  110. ```sh
  111. moi openclaw init
  112. moi openclaw init <agent> # only needed when you have multiple agents
  113. moi start # or run Moi as a service
  114. ```
  115. Your agent's workspace will show up at `http://localhost:13337`.
  116. ## Hermes
  117. [Hermes](https://hermes-agent.nousresearch.com/) is a self-hosted personal
  118. agent from Nous Research.
  119. Moi connects each Hermes profile to one workspace. A profile is a separate
  120. agent identity with its own model, keys, persona, skills, and memories.
  121. Run `moi hermes init` to connect your profile. If Moi finds more than one, it
  122. lists them so you can rerun the command with the profile you want:
  123. ```sh
  124. moi hermes init
  125. moi hermes init <profile> # only needed when you have multiple profiles
  126. moi start # or run Moi as a service
  127. ```
  128. The profile's workspace appears at `http://localhost:13337`. To add another,
  129. create it with `hermes profile create <name>`, then run `moi hermes init` again.
  130. Moi connects to Hermes through its built-in
  131. [Agent Client Protocol](https://agentclientprotocol.com/) server. It starts a
  132. `hermes -p <profile> acp` process and communicates over stdio, so there is no
  133. gateway, port, or additional API key to configure.
  134. # Manage Moi
  135. ## Run Moi as a service
  136. Instead of keeping `moi start` in a terminal, install Moi as a user-level
  137. service. It starts on login, restarts on crash, and survives reboots
  138. (launchd on macOS, a systemd user unit on Linux; no root needed):
  139. ```sh
  140. moi service install # install and start
  141. moi service # state, unit path, server version
  142. moi service logs -f # follow server logs
  143. moi service restart
  144. moi service uninstall
  145. ```
  146. Anything your agents need belongs in the workspace env (`moi env set` or
  147. `.env`). To capture shell variables instead, run
  148. `moi service install --env MY_TOKEN,OTHER`.
  149. Rerun `moi service install` after changing captured vars (or moving bun). On
  150. macOS the first install may show a "background item added" notification for
  151. "moi"; that is the Moi service. On headless Linux, lingering is enabled
  152. automatically when possible so the service outlives your SSH session.
  153. ## Update Moi
  154. ```sh
  155. moi update # update to the latest release
  156. moi update --check # only check; exit 0 up to date, 1 update available, 2 check failed
  157. ```
  158. `moi update` checks npm for the latest release and updates through whichever
  159. package manager owns the install (bun, npm, pnpm, or yarn). A service-managed
  160. server is restarted onto the new version; a foreground `moi start` only gets
  161. a warning, so restart it yourself. Prerelease installs (`…-next.N`) are left
  162. alone. `--check` changes nothing and is made for scripts and agents.
  163. `moi status` shows when the running server and CLI versions differ, however
  164. the update happened.
  165. # No crypto token
  166. Moi has **no** cryptocurrency, token, coin, or NFT, official or otherwise, on
  167. any chain, and never will. Any token using the Moi name, logo, or the
  168. maintainer's name (on pump.fun, Solana, or anywhere else) is an unauthorized
  169. scam with no affiliation to, endorsement from, or benefit to this project.
  170. The maintainer will never announce, endorse, or accept proceeds from a token.
  171. Don't buy it, and report it to the platform where it's listed.
  172. License: [Elastic 2.0](LICENSE).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.