Skip to content
Back to search

clauderipple

Use GPT (ChatGPT Plus/Pro), DeepSeek, Kimi, Grok and 400+ models in the Claude Desktop Code tab and Claude Code while staying signed in to your Claude subscription — no third-party gateway (3P) mode. Several ChatGPT accounts that take over when one runs out; Claude inside Codex too. Not affiliated with Anthropic or OpenAI.

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/PBJ-2/clauderipple 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/PBJ-2/clauderipple'

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/pbj-2--clauderipple/badge.svg)](/listings/pbj-2--clauderipple)

HTML

<a href="/listings/pbj-2--clauderipple"><img src="/api/listings/pbj-2--clauderipple/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.

These steps run a script from the network that findsafeskills does not scan.

**macOS**

curl -fsSL https://raw.githubusercontent.com/PBJ-2/clauderipple/main/scripts/install.sh | sh

**Windows** (PowerShell)

irm https://raw.githubusercontent.com/PBJ-2/clauderipple/main/scripts/install.ps1 | iex

**Linux** (a desktop session with systemd)

curl -fsSL https://raw.githubusercontent.com/PBJ-2/clauderipple/main/scripts/install.sh | sh

The router runs as a per-user systemd service: `systemctl --user status clauderipple`, and `journalctl --user -u clauderipple` for what it printed. Picker mode trusts the local CA in the NSS database Chromium reads (`~/.pki/nssdb`, your user only, no sudo), which takes `certutil`: run `sudo apt install libnss3-tools` (Fedora: `sudo dnf install nss-tools`) before `clauderipple picker on`.

That is the whole prerequisite list. The script uses a Node 24+ you already have, and downloads the official build into `~/.clauderipple/runtime` when you have none, verifying it against the checksum nodejs.org publishes. Then it sets ClaudeRipple up: a local certificate, two lines in `~/.claude/settings.json`, and a background router that starts with your computer. No administrator rights, no password.

With Node already installed you can skip the script:

npm install -g clauderipple
clauderipple install

Then open the dashboard and add a provider:

clauderipple ui

**Providers** → add ChatGPT or paste an API key → **Model mapping**.

**The menu-bar / tray app** is optional. It shows the router's state and opens the dashboard in a click:

clauderipple tray

It runs on Electron, which is about 270MB and is therefore not installed by default. `clauderipple tray --install` fetches it once; everything else works without it.

**Updating.** The tray offers a new version when there is one (**Check for Updates…**), or:

clauderipple update

It updates the way you installed — the install script again, or npm — and restarts the router on the new version. The tray's Electron is kept across updates.

Optional, for real names in the Desktop picker: **Clients → Claude Desktop → Model picker → on**, then quit and reopen Claude Desktop. To trust the local certificate for your user only, macOS asks for your login password and Windows shows a confirmation dialog with the fingerprint; Linux adds it to your NSS database without asking. ClaudeRipple never sees a password, and no platform needs administrator rights.

> **Closing Claude Desktop's window is not enough.** It keeps running, and the next > launch reuses it without reading the new setting. Quit it properly (macOS: ⌘Q; > Windows: the tray icon, or Task Manager; Linux: Ctrl+Q or the tray icon, until > `pgrep -f claude-desktop` prints nothing) or the picker will silently not change.

<details> <summary>From source (Node 24)</summary>

git clone https://github.com/PBJ-2/clauderipple && cd clauderipple && npm install
node packages/cli/src/index.ts install   # certs, settings.json env, supervisor (launchd / Task Scheduler), end-to-end probe
node packages/cli/src/index.ts ui        # open the local GUI in your browser

`uninstall` reverses everything and restores `~/.claude/settings.json` from a backup. Other commands: `status`, `start`, `stop`, `restart`, `logs -f`, `login`, `logout`, `claude-login`, `claude-logout`, `picker on|off`, `agent-title on|off`, `mod on|off`, `codex on|off`, `update`, `image "<prompt>"`. </details>

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 · README · 743 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. <p align="center">
  2. <picture>
  3. <source media="(prefers-color-scheme: dark)" srcset="brand/mark-dark.svg">
  4. <source media="(prefers-color-scheme: light)" srcset="brand/mark.svg">
  5. <img src="brand/mark.svg" width="120" height="120" alt="The ClaudeRipple mascot: a hollow ripple ring with a curling wave crest">
  6. </picture>
  7. </p>
  8. <h1 align="center">ClaudeRipple</h1>
  9. <p align="center"><b>English</b> · <a href="README.ko.md">한국어</a> · <a href="README.zh-CN.md">简体中文</a> · <a href="README.ja.md">日本語</a> · <a href="README.es.md">Español</a> · <a href="README.pt-BR.md">Português (BR)</a></p>
  10. <p align="center">
  11. <b>Chat stays Claude. Only Code becomes GPT.</b><br>
  12. Claude Desktop's own setting for other models switches the <b>whole app</b> — chat, phone Remote
  13. Control and connectors go with it.<br>ClaudeRipple turns nothing off: you stay signed in to your
  14. Claude subscription while GPT, DeepSeek, Kimi, Grok<br>or 400+ other models answer in the
  15. <b>Code tab</b>. Claude inside the <b>Codex app</b> and <b>Codex CLI</b> too.
  16. </p>
  17. <p align="center">
  18. <a href="https://apps.apple.com/us/app/id6823063731"><img height="40" alt="App Store" src="https://img.shields.io/badge/App%20Store-Download-0D96F6?style=for-the-badge&logo=apple&logoColor=white"></a>
  19. <a href="https://play.google.com/store/apps/details?id=com.clauderipple.remote&hl=en"><img height="40" alt="Google Play" src="https://img.shields.io/badge/Google%20Play-Download-01875F?style=for-the-badge&logo=googleplay&logoColor=white"></a>
  20. <a href="https://discord.gg/R8ZgGSNHjw"><img height="40" alt="Discord" src="https://img.shields.io/badge/Discord-Join-5865F2?style=for-the-badge&logo=discord&logoColor=white"></a>
  21. </p>
  22. <p align="center">
  23. <a href="https://www.npmjs.com/package/clauderipple"><img alt="npm" src="https://img.shields.io/npm/v/clauderipple?label=npm"></a>
  24. <img alt="Alpha" src="https://img.shields.io/badge/status-alpha-orange">
  25. <a href="LICENSE"><img alt="GPL-3.0" src="https://img.shields.io/badge/license-GPL--3.0-blue"></a>
  26. <img alt="macOS" src="https://img.shields.io/badge/macOS-arm64%20%7C%20x64-black">
  27. <img alt="Windows" src="https://img.shields.io/badge/Windows-arm64%20%7C%20x64-0078D4">
  28. <img alt="Linux" src="https://img.shields.io/badge/Linux-x64%20(systemd)-FCC624">
  29. <img alt="Node" src="https://img.shields.io/badge/node-24%2B-success">
  30. <a href="README.ko.md"><img alt="한국어" src="https://img.shields.io/badge/docs-한국어-red"></a>
  31. </p>
  32. <p align="center">
  33. <a href="https://github.com/PBJ-2/clauderipple/releases/download/v0.9.0/clauderipple-demo.mp4"><img src="docs/media/demo.webp" width="880" alt="A one-minute tour: other models in Claude Desktop's own model menu, effort set for every prompt, and each subagent sent to the model that fits. Click for the full video."></a><br>
  34. <sub>A one-minute tour: other models in Claude Desktop's own model menu, effort set for every prompt, and each subagent sent to the model that fits. Click for the full video.</sub>
  35. </p>
  36. <table align="center">
  37. <tr>
  38. <td align="center" width="34%"><img src="docs/media/picker.png" width="300" alt="GPT, DeepSeek, GLM and Gemini in Claude Desktop's own model menu"><br><sub>GPT, DeepSeek, GLM and Gemini in Claude Desktop's own model menu</sub></td>
  39. <td align="center" width="66%"><img src="docs/media/answer.png" alt="GPT-6 Astra working in the Code tab, every request logged on the right"><br><sub>GPT-6 Astra working in the Code tab, every request logged on the right</sub></td>
  40. </tr>
  41. <tr>
  42. <td colspan="2" align="center"><img src="docs/media/subagents.png" alt="Subagents on DeepSeek and GLM, named in the background-task panel"><br><sub>Subagents on DeepSeek and GLM, named in the background-task panel</sub></td>
  43. </tr>
  44. <tr>
  45. <td colspan="2" align="center"><img src="docs/media/mod-band.png" alt="The ripple mod: effort, context, cache timer, and which model answered with its cache hit"><br><sub>The ripple mod: effort, context, cache timer, and which model answered with its cache hit</sub></td>
  46. </tr>
  47. <tr>
  48. <td align="center" width="50%"><img src="docs/media/mod-alerts.png" alt="Warnings when the cache goes cold or a limit is near, and a compaction offer"><br><sub>Warnings when the cache goes cold or a limit is near, and a compaction offer</sub></td>
  49. <td align="center" width="50%"><img src="docs/media/mod-settings.png" alt="Quick settings: automatic effort, the judge, worker modes"><br><sub>Quick settings: automatic effort, the judge, worker modes</sub></td>
  50. </tr>
  51. </table>
  52. ---
  53. ## 1P, not 3P: the difference this whole project exists for
  54. Claude Desktop already ships an official way to use other models — the **inference
  55. gateway** setting, what the app calls third-party or **3P** mode. Turning it on is
  56. not a per-session choice. It switches the **whole app** into a different deployment
  57. mode at launch, and that mode is a different product:
  58. - The window stops loading `claude.ai` and loads a local bundle instead. claude.ai
  59. `/api/` and `/v1/` calls answer `custom_3p_not_available` 503.
  60. - "Chat" is no longer claude.ai chat. It is a Claude Code local-agent session.
  61. - Remote Control and side sessions are switched off outright
  62. (`shouldEnableSessionsBridge()` returns false).
  63. - Anthropic's own claude.ai connectors, Claude Design, mobile continuity and chat
  64. search go with it.
  65. There is no middle setting. "Chat on Claude, Code on the gateway" is impossible for
  66. anyone, because 1P mode routes nothing through the gateway at all. (All of this is
  67. read out of the app's own bundle; see [docs/ARCHITECTURE.md](docs/ARCHITECTURE.md) §2.)
  68. **ClaudeRipple never touches that setting.** The app stays signed in to your Claude
  69. subscription in 1P mode, and another model answers in the Code tab and its subagents
  70. — under its real name in the model picker, at the reasoning effort you chose. You do
  71. not trade your Claude account for a GPT one. You keep both, in one app.
  72. Three things follow from staying in 1P, and no gateway-based tool can offer any of them:
  73. - **Code from your phone, on GPT.** Remote Control is off in 3P. It is on here, so you
  74. can pick a model and send a task from a phone and watch the commits land.
  75. - **Connectors alive while another model works.** 3P stubs claude.ai calls, which takes
  76. Anthropic's own connectors with them. Read a Google Drive doc and have GPT write the code.
  77. - **Cloud and side sessions.** Gone in 3P. Untouched here.
  78. If you pay for Claude, 3P is not an option — it throws away the half of the
  79. subscription you are paying for. This is the only way to keep it and still choose
  80. the model.
  81. ## FAQ
  82. ### Can I use GPT in the Claude Desktop app?
  83. Yes — in the **Code tab** and its subagents. Install ClaudeRipple, sign in with a
  84. ChatGPT Plus or Pro subscription, and GPT (GPT-6 Sol, Luna, Astra, GPT-5.6 Terra and
  85. the rest) shows up under its real name in the model picker, with the reasoning effort
  86. you pick. The app itself is not modified and not switched into any other mode. The
  87. general **chat** tab stays Claude: no tool can route it without turning the whole app
  88. into gateway mode (see [1P, not 3P](#1p-not-3p-the-difference-this-whole-project-exists-for)).
  89. ### How do I use GPT in Claude Code?
  90. The same install covers the terminal `claude` CLI. ClaudeRipple is a local proxy that
  91. Claude Code is pointed at through two lines in `~/.claude/settings.json`, so your
  92. skills, hooks, MCP servers, `CLAUDE.md` and subagents keep working while another model
  93. answers.
  94. ### Do I have to turn on Claude Desktop's "third-party inference" (gateway) setting?
  95. No — and you should not. That setting switches the whole app into 3P mode, which stops
  96. loading claude.ai, replaces chat with a local-agent session and turns Remote Control
  97. off. ClaudeRipple never touches it.
  98. ### Will I lose claude.ai chat, Remote Control or my connectors?
  99. No. The app stays signed in to your Claude subscription, so chat, phone Remote Control,
  100. cloud sessions and Anthropic's own connectors keep working. That is the whole point of
  101. staying in 1P.
  102. ### Can I still use Claude itself?
  103. Yes. Model mapping is per model name: leave a name mapped to Claude and it answers as
  104. Claude. Most people map one or two names to GPT and leave the rest alone.
  105. ### Do I need an OpenAI API key?
  106. No. Any ChatGPT plan that includes Codex works — Plus, Pro and Business, and Enterprise or
  107. Edu where the workspace admin has turned Codex on. You sign in from the app and no API key
  108. is involved. The plan's usage limits apply; once they are used up, Codex credits you have
  109. bought carry on. API keys are for the other providers (DeepSeek, Kimi, GLM, OpenRouter and
  110. anything else with an Anthropic- or OpenAI-compatible endpoint).
  111. ### Which models can I use?
  112. GPT through a ChatGPT subscription, and DeepSeek, Kimi, GLM, Grok, Qwen and 400+ more
  113. through OpenRouter or a direct API key. Claude models keep working as they are.
  114. ### Does it work on Windows?
  115. Yes, on Windows and macOS, arm64 and x64. Install, dashboard, restart and uninstall are
  116. measured on Windows 11 arm64; the x64 runtime is measured under emulation. Linux works too
  117. where there is a systemd user session: install, a Code-tab request through the router, a
  118. restart in the middle of a streaming answer and uninstall are measured on Ubuntu 24.04 x64
  119. with Claude Desktop 2.2553.13, and so is picker mode: models from `cli.extraModels` show
  120. in the Code-tab picker next to Claude's own.
  121. ### Is my code sent anywhere else?
  122. No. The proxy runs on your own machine. Requests go to the provider you configured and
  123. nowhere else, and credentials stay in your home directory. See [Privacy](#privacy).
  124. ### Will my Claude account get banned? Is this against Anthropic's terms?
  125. What ClaudeRipple does with Claude is small and checkable: a request that stays on Claude
  126. leaves your machine byte for byte as Claude Code sent it, and a request routed to GPT or
  127. another provider never reaches Anthropic at all. It works through the proxy and
  128. certificate settings Claude Code documents for corporate networks (`HTTPS_PROXY`,
  129. `NODE_EXTRA_CA_CERTS` — [Enterprise network configuration](https://code.claude.com/docs/en/corporate-proxy)),
  130. not by patching the app. The reports of banned accounts are about
  131. the opposite direction — a Claude subscription login used from **another** program.
  132. ClaudeRipple has one optional feature of that kind: Claude inside Codex using your
  133. subscription login. That use is subject to Anthropic's terms; an Anthropic API key avoids
  134. the question. Rotating several Claude subscriptions is likewise your call under those
  135. terms. Only Anthropic can say how an account is treated, so nothing here is a promise.
  136. ### What about Gemini through a Google account?
  137. An AI Studio API key is the way Google supports, and the safe one. The Google account
  138. sign-in (Antigravity) is different: Google's [Antigravity terms](https://antigravity.google/terms)
  139. (section 6) and the [Gemini CLI FAQ](https://geminicli.com/docs/resources/faq/) name using
  140. those credentials from another program as a breach. In February 2026 Google suspended
  141. accounts used this way — the suspension also cut off Gemini CLI and Code Assist — and
  142. [said](https://github.com/google-gemini/gemini-cli/discussions/20632) a second violation
  143. is permanent. ClaudeRipple offers the sign-in because others do and people asked, shows
  144. this warning before it starts, and does nothing to hide the traffic. Use a secondary
  145. account if you use it at all.
  146. ### Can I use several ChatGPT accounts?
  147. Yes. Each `clauderipple login` (or **+ Add ChatGPT account** in the dashboard) adds one.
  148. When an account hits its usage limit, the same request moves to the next account and the
  149. spent one rests until its window resets; a conversation stays on the account that
  150. answered, to keep its prompt cache. This applies to the Claude Desktop Code tab, Claude
  151. Code, and GPT in the Codex app and CLI.
  152. ### Can I use Claude inside Codex?
  153. Yes. `clauderipple codex on` adds ClaudeRipple as a provider, and Claude models appear in
  154. the Codex app's and CLI's model list under their own names. They answer through your
  155. Claude Code login or an Anthropic API key. See [Codex app and Codex CLI](#codex-app-and-codex-cli).
  156. ### Can it generate images?
  157. Yes, with a ChatGPT subscription and no API key:
  158. `clauderipple image "a red paper boat" -o boat.png --aspect square`. `--ref FILE` adds
  159. a reference image and `--transparent` a transparent background. One image takes about
  160. 30 seconds, and the subscription chooses the resolution and quality. Subagents
  161. ClaudeRipple generates are told the same call, so they can make images too. See
  162. [Claude Code](#claude-code-terminal-remote-control-subagents).
  163. ### Does GPT do worse inside Claude Code?
  164. Claude Code's system prompt and tool descriptions are written for Claude, so another model
  165. reads instructions tuned for someone else, and it may behave a little differently than in
  166. its own client. ClaudeRipple tells each model its own name and effort, reshapes tool
  167. schemas a vendor would reject, and lets you append instructions per provider. A common
  168. setup is Claude as the main model with GPT or DeepSeek as subagents.
  169. ### How is this different from CC Switch, opencodex or claude-code-router?
  170. See the [comparison table](#one-tool-instead-of-four). In short: ClaudeRipple reaches
  171. the Claude **Desktop** Code tab while the app stays signed in to Claude, and shows other
  172. models in its picker under their real names. CC Switch also reaches the desktop app, but
  173. by turning on its gateway (3P) mode.
  174. ### Is it free?
  175. Yes. ClaudeRipple is open source under GPL-3.0. You pay only for the subscriptions or API
  176. keys you already use.
  177. ## One tool instead of four
  178. Similar tools let the Claude Code CLI or Codex CLI use other models, but they either
  179. cannot reach the Claude **desktop app** or reach it only through its gateway (3P) mode,
  180. and the moment you switch models you lose the Claude-subscription side (claude.ai chat,
  181. Remote Control, cloud sessions). ClaudeRipple covers the desktop app, the terminal and
  182. Codex from one menu-bar app, runs both subscriptions side by side, and keeps the
  183. Claude Code harness intact: your skills, hooks, MCP servers, `CLAUDE.md`, subagents
  184. and claude.ai connectors keep working while another model does the thinking.
  185. | | ClaudeRipple | CC Switch (checked 2026-09-24) | opencodex / openclaude (checked 2026-09-16) | claude-code-router | Claude Desktop gateway (3P) mode |
  186. |---|---|---|---|---|---|
  187. | Claude **Desktop** Code tab **without** gateway (3P) mode | ✅ | ❌ ² | ❌ ¹ | ❌ | ❌ by definition |
  188. | Keeps claude.ai chat, Remote Control, cloud sessions, connectors | ✅ | ❌ | ❌ | ❌ | ❌ |
  189. | Claude and GPT subscriptions side by side | ✅ | ❌ one provider at a time | ❌ all-or-nothing | ❌ | ❌ |
  190. | Real model names in the Desktop picker | ✅ | ❌ `claude-*` role names ² | ❌ | ❌ | partial |
  191. | Several ChatGPT accounts, switching when one runs out (Desktop and Codex) | ✅ | ❌ switched by hand ³ | ✅ | ❌ | ❌ |
  192. | Terminal `claude` CLI | ✅ | ✅ | ✅ | ✅ | ✅ |
  193. | Codex **app** and Codex CLI → Claude | ✅ | ✅ | ✅ | ❌ | ❌ |
  194. | Prompt cache on translated providers | **94–99 %** measured | not measured | not measured | varies | n/a |
  195. | Subagents named by real model in the task panel | ✅ | ❌ | ❌ | ❌ | ❌ |
  196. | Settings GUI, no terminal needed | ✅ | ✅ | ❌ | ❌ | ❌ |
  197. | Signed, notarized app with its own runtime | ✅ | ✅ | ❌ | ❌ | – |
  198. ¹ opencodex's README showed Claude Desktop in a demo but published no setup steps for it,
  199. in the repository or on its documentation site (checked 2026-09-16). At the time the only
  200. public way into the desktop app was the official gateway setting — the last column.
  201. ² CC Switch reaches Claude Desktop by writing the app's third-party profile
  202. (`Claude-3p/claude_desktop_config.json`, `"inferenceProvider": "gateway"`), which is the
  203. last column. Through its local gateway the picker shows only `claude-sonnet-*`,
  204. `claude-opus-*` and `claude-haiku-*` role names ([its manual](https://github.com/farion1231/cc-switch/blob/main/docs/user-manual/en/2-providers/2.6-claude-desktop.md)).
  205. CC Switch manages more tools than this table covers (Gemini CLI, OpenCode and others) and
  206. syncs MCP servers and skills across them.
  207. ³ Each ChatGPT account is bound to its own provider card, and official ChatGPT cards are
  208. kept out of the failover queue (v3.20.0 release notes).
  209. The desktop app's own "third-party inference" setting flips the whole app into another
  210. mode: you lose claude.ai chat, Remote Control and cloud sessions. Tools that replace
  211. `ANTHROPIC_BASE_URL` never reach the desktop app, and the ones that do reach it write
  212. that same setting. ClaudeRipple instead is a
  213. tiny HTTPS proxy that only the Claude Code process trusts: requests for the models
  214. you map go to your provider, everything else goes to Anthropic byte for byte.
  215. ## What you get
  216. - **Any model in Claude Desktop and Claude Code.** GPT-6 Sol / Luna / Astra and
  217. GPT-5.6 Terra / Sol / Luna through your ChatGPT Plus/Pro subscription (new models
  218. appear the day OpenAI serves them), Google Gemini, or DeepSeek, Kimi, GLM,
  219. MiniMax, Qwen, Grok, Mistral, Groq, Together, Fireworks, OpenRouter (400+ models)
  220. and local Ollama / LM Studio. Under their real names in the picker, or mapped onto
  221. a Claude name.
  222. - **Claude in the Codex app and Codex CLI.** A local OpenAI-compatible endpoint that
  223. Codex treats as a provider; Claude models appear in Codex's own model list. Uses
  224. your current Claude Code login or an Anthropic API key.
  225. - **Multiple Claude accounts without breaking the prompt cache.** Opt a native Claude
  226. provider into account rotation, add subscriptions from the dashboard, and each
  227. conversation stays on the account that answered. Before the response starts, a
  228. quota or authentication refusal moves that turn to the next account. This native
  229. Claude Desktop/Code path is separate from the translated Codex ingress, which picks
  230. one available login and does not rotate accounts.
  231. - **Several ChatGPT accounts.** Each `clauderipple login` (or **+ Add ChatGPT account**
  232. in the dashboard) adds one. When an account hits its limit, the next one takes over
  233. the same request and the spent one rests until its window resets; a conversation
  234. stays on the account that answered, to keep its cache. This covers the Claude Desktop
  235. Code tab and subagents **and GPT in the Codex app and CLI** — Codex keeps its own
  236. sign-in and moves to the next account without signing out.
  237. - **The full Claude Code harness, untouched.** Skills, hooks, MCP, `CLAUDE.md`,
  238. subagents, plan mode, Remote Control on your phone: nothing is turned off.
  239. - **Correct by construction.** Prompt caching preserved (Anthropic cache breakpoints
  240. and stable OpenAI prefixes), Claude Code's server-side threads handled, tool calls
  241. and images round-tripped, reasoning effort clamped to what each model accepts,
  242. Anthropic-only request fields stripped for compatible vendors.
  243. - **A real request log.** Who asked, which model answered, tokens in / cached / out,
  244. latency, status, per request, with a one-hour summary.
  245. - **Subagent name tags.** The background-task panel shows `Terra·high · Review`
  246. instead of a generic "Agent".
  247. - **The ripple mod, above the prompt.** A band shows the reasoning effort, context, cache
  248. timer, and which model answered your last request with its cache hit. A small judge
  249. model (Haiku 5.5 or Jev) reads each prompt with the recent turns and sets the effort,
  250. low through max. Delegated work goes to `ripple:auto`, and the judge sends each task to
  251. a worker by difficulty: light work to DeepSeek or GLM, hard work to GPT or Opus. Cards
  252. warn when the cache has gone cold, the context gets heavy or a plan window passes 80%,
  253. and offer a compaction in place. `Log` (or
  254. `/ripple-log`) opens the request log in a side pane. One switch on the Clients screen,
  255. or `clauderipple mod on`.
  256. - **Images from your ChatGPT subscription.** `clauderipple image "<prompt>"`, with
  257. reference images and transparent backgrounds. Subagents are told how, so a GPT or
  258. DeepSeek worker can make an image and look at it without an image tool.
  259. - **Close the laptop, Claude keeps working.** On macOS and Windows the laptop stays awake
  260. with the lid shut while Claude works (and 10 minutes after), so a task you started from
  261. your phone finishes. If the Wi-Fi loses its internet, it moves to your phone's hotspot
  262. (iPhone or Android) and back to a known Wi-Fi later. Off until you turn it on in the
  263. dashboard's Laptop view; see [Lid closed and the phone hotspot](#lid-closed-and-the-phone-hotspot).
  264. - **Lean workers, if you want them.** One switch gives generated subagents only the
  265. tools a worker uses instead of every MCP server and plugin of the session: one
  266. worker's first request fell from 33k to 9k tokens.
  267. - **Made for people who don't want a terminal.** Provider presets with one-click
  268. connection tests and model discovery, drop-down model mapping, auto-save, a UI in
  269. English, Korean, Chinese, Japanese, Spanish and Portuguese. A menu-bar app that carries its own runtime and sets itself up on
  270. first launch. Signed and notarized.
  271. <p align="center">
  272. <img src="docs/media/mapping.png" width="880" alt="Model mapping: which model answers for each Claude name, with per-model reasoning effort">
  273. </p>
  274. ## Ripple Remote: every model, from your phone, for hours
  275. Start a job in the Claude desktop app, close the laptop, and keep going from the couch.
  276. **Ripple Remote** is the phone app for ClaudeRipple: the same sessions, whichever model
  277. runs them, on your phone, and it does not drop you after 30 minutes.
  278. <p align="center">
  279. <img src="docs/media/remote/store/en-1.png" width="260" alt="A remote that doesn't time out: a GPT session followed for 1h 47m">
  280. <img src="docs/media/remote/store/en-2.png" width="260" alt="GPT, DeepSeek and Gemini in one model picker on the phone">
  281. <img src="docs/media/remote/store/en-3.png" width="260" alt="Approve a DeepSeek session's request from the phone">
  282. </p>
  283. - **Hours, not minutes.** The connection holds as long as your computer is on. Watch a
  284. two-hour refactor through to the end and approve it when it asks.
  285. - **Any model you route.** GPT, DeepSeek, Gemini or Claude: every session looks and works
  286. the same. Switch the next turn's model and reasoning effort from your phone.
  287. - **Your Claude app sessions, as they are.** Sessions you started in the Code tab on your
  288. Mac continue on the phone — the live timeline, tool calls, diffs, subagents and their
  289. models, plan approvals and question cards, photos and files.
  290. - **Your Mac's status bar in your pocket.** The model that answered, the cache and your
  291. usage limits, as the ClaudeRipple band shows them on the Mac.
  292. - **Pairs with the laptop features.** With [lid-closed awake and the phone
  293. hotspot](#lid-closed-and-the-phone-hotspot), the laptop keeps working in a bag and stays
  294. online when the Wi-Fi does not.
  295. - **Private by design.** Messages, attachments and notifications are end-to-end
  296. encrypted; the relay and the push service cannot read them. No account, no ads, no
  297. tracking, free.
  298. Install ClaudeRipple on your computer, open the dashboard's **Phone** view and scan the
  299. QR code once. iPhone: [App Store](https://apps.apple.com/us/app/id6823063731). Android:
  300. [Google Play](https://play.google.com/store/apps/details?id=com.clauderipple.remote&hl=en).
  301. Claude app sessions are supported on macOS; the app has a demo you can try without a
  302. computer.
  303. ## Roadmap
  304. The full plan, with the issues behind each item, is in [docs/ROADMAP.md](docs/ROADMAP.md). In order:
  305. 1. **See where your limit goes.** Why a request missed the cache, what each request is made of, a limit forecast with alerts, task-finished alerts, a guard against dangerous commands, and image descriptions for models that cannot read images.
  306. 2. **Settings in one place.** One page for plugins, MCP servers, skills and hooks, with the tokens each one really costs; popular bundles in one click; one rules file for Claude and Codex.
  307. 3. **Memory.** Every session distilled daily into decisions, records and open items, and only the few that matter brought back when you start a session, ask about it, or are about to edit a file.
  308. 4. **Accounts handled for you.** Use first the account whose weekly limit resets soonest; hold a request when every account is out and send it when one resets; import accounts from other tools.
  309. 5. **Save tokens, measured honestly.** A before-and-after cost meter first, then savings in the router that do not break the cache.
  310. Ideas and votes are welcome in [issues](https://github.com/PBJ-2/clauderipple/issues).
  311. ## Install
  312. **macOS**
  313. ```sh
  314. curl -fsSL https://raw.githubusercontent.com/PBJ-2/clauderipple/main/scripts/install.sh | sh
  315. ```
  316. **Windows** (PowerShell)
  317. ```powershell
  318. irm https://raw.githubusercontent.com/PBJ-2/clauderipple/main/scripts/install.ps1 | iex
  319. ```
  320. **Linux** (a desktop session with systemd)
  321. ```sh
  322. curl -fsSL https://raw.githubusercontent.com/PBJ-2/clauderipple/main/scripts/install.sh | sh
  323. ```
  324. The router runs as a per-user systemd service: `systemctl --user status clauderipple`, and
  325. `journalctl --user -u clauderipple` for what it printed. Picker mode trusts the local CA in the
  326. NSS database Chromium reads (`~/.pki/nssdb`, your user only, no sudo), which takes `certutil`:
  327. run `sudo apt install libnss3-tools` (Fedora: `sudo dnf install nss-tools`) before `clauderipple picker on`.
  328. That is the whole prerequisite list. The script uses a Node 24+ you already have, and
  329. downloads the official build into `~/.clauderipple/runtime` when you have none,
  330. verifying it against the checksum nodejs.org publishes. Then it sets ClaudeRipple up:
  331. a local certificate, two lines in `~/.claude/settings.json`, and a background router
  332. that starts with your computer. No administrator rights, no password.
  333. With Node already installed you can skip the script:
  334. ```sh
  335. npm install -g clauderipple
  336. clauderipple install
  337. ```
  338. On Windows, PowerShell's default execution policy refuses the `clauderipple.ps1` npm adds, so
  339. run `clauderipple.cmd install` there (or use Command Prompt). `install` removes that file, and
  340. plain `clauderipple` works in PowerShell from then on. The install script never leaves one.
  341. Then open the dashboard and add a provider:
  342. ```sh
  343. clauderipple ui
  344. ```
  345. **Providers** → add ChatGPT or paste an API key → **Model mapping**.
  346. **The menu-bar / tray app** shows the router's state, opens the dashboard in a click, and
  347. keeps a repair screen that works while the router is down:
  348. ```sh
  349. clauderipple tray
  350. ```
  351. On macOS and Windows the install script adds it by default. It runs on Electron, about 270MB,
  352. fetched once; set `CLAUDERIPPLE_NO_TRAY=1` before the install command to skip it. On Linux, or
  353. after an npm-only install, `clauderipple tray --install` adds it. Everything else works without it.
  354. **Updating.** The tray offers a new version when there is one (**Check for Updates…**), or:
  355. ```sh
  356. clauderipple update
  357. ```
  358. It updates the way you installed — the install script again, or npm — and restarts the router
  359. on the new version. The tray's Electron is kept across updates.
  360. Optional, for real names in the Desktop picker: **Clients → Claude Desktop → Model
  361. picker → on**, then quit and reopen Claude Desktop. To trust the local certificate
  362. for your user only, macOS asks for your login password and Windows shows a
  363. confirmation dialog with the fingerprint; Linux adds it to your NSS database without asking.
  364. ClaudeRipple never sees a password, and no platform needs administrator rights.
  365. > **Closing Claude Desktop's window is not enough.** It keeps running, and the next
  366. > launch reuses it without reading the new setting. Quit it properly (macOS: ⌘Q;
  367. > Windows: the tray icon, or Task Manager; Linux: Ctrl+Q or the tray icon, until
  368. > `pgrep -f claude-desktop` prints nothing) or the picker will silently not change.
  369. <details>
  370. <summary>From source (Node 24)</summary>
  371. ```bash
  372. git clone https://github.com/PBJ-2/clauderipple && cd clauderipple && npm install
  373. node packages/cli/src/index.ts install # certs, settings.json env, supervisor (launchd / Task Scheduler), end-to-end probe
  374. node packages/cli/src/index.ts ui # open the local GUI in your browser
  375. ```
  376. `uninstall` reverses everything and restores `~/.claude/settings.json` from a backup.
  377. Other commands: `status`, `start`, `stop`, `restart`, `logs -f`, `login`, `logout`,
  378. `claude-login`, `claude-logout`, `picker on|off`, `agent-title on|off`, `mod on|off`,
  379. `codex on|off`, `update`, `image "<prompt>"`.
  380. </details>
  381. ## Clients
  382. ### Claude Desktop
  383. Works out of the box after install. Map models on **Model mapping** (auto-saves),
  384. or turn on **picker mode** to see provider models by name in the app's own picker.
  385. Effort chosen in the app is passed through; unsupported levels are clamped.
  386. ### Claude Code (terminal, Remote Control, subagents)
  387. Same router, same mapping. `/model gpt-5.6-terra` lists the models you added.
  388. Subagents follow the routing; a `[[gpt: sol@xhigh]]` marker in a subagent prompt
  389. overrides the model for that call, and the task panel shows the real model name.
  390. **You do not have to build an agent harness for this.** Point the subagent slot at
  391. a routed model and every subagent runs on it — no agent file, nothing else:
  392. ```jsonc
  393. // Settings → CLI models, or "cli": { "models": { … } } in the config
  394. { "subagent": "gpt-5.6-terra" } // → CLAUDE_CODE_SUBAGENT_MODEL
  395. ```
  396. **Every model you tick is also a named subagent.** The router writes
  397. `~/.claude/agents/<name>.md` for each model exactly one provider carries
  398. (`gpt-5.6-terra` → `gpt-5-6-terra`, `deepseek-v4.1-flash` → `deepseek-v4-1-flash`),
  399. so "have DeepSeek do it" needs nothing beyond ticking the model: the Agent tool
  400. lists it, `subagent_type: "deepseek-v4-1-flash"` runs it, and the task panel
  401. names it. Untick the model and its file goes away. The router only ever touches
  402. the files it wrote (recorded in `generated-agents.json`); an agent file you wrote
  403. yourself under the same name wins and is left alone, so a custom brief is still
  404. just Claude Code's own agent file with a `model:` the router knows:
  405. ```markdown
  406. ---
  407. name: reviewer
  408. description: Independent review on GPT-5.6 Sol.
  409. model: gpt-5.6-sol@medium
  410. ---
  411. You are the reviewer for this session. Verify the change yourself and report
  412. the conclusion only.
  413. ```
  414. Per-call effort goes on the **first line** of the subagent prompt:
  415. `[[ripple: gpt-5-6-terra@high]]` — any agent name or model id works there. A
  416. marker anywhere else is prose and is ignored (a compaction summary that quotes
  417. one must not re-route the session). A model the router cannot place is refused
  418. by name (`400 ClaudeRipple: no provider declares "…"`) instead of being sent on
  419. to fail somewhere less legible; native Claude models always pass through.
  420. Set `"cli": { "agentFiles": false }` to turn generation off.
  421. By default a generated worker inherits every tool of the session, so each MCP
  422. server and plugin rides along on every worker request. `"cli": {
  423. "limitWorkerTools": true }` (or **Worker tools** in the dashboard) gives it only
  424. the read, edit, run, search and skill tools, which cut one worker's first
  425. request from about 33k to 9k tokens. Such a worker has no browser or simulator
  426. tools. A worker that needs them belongs in an agent file of your own, which is
  427. never touched.
  428. With a ChatGPT provider configured, images come from the same subscription:
  429. ```bash
  430. clauderipple image "a red paper boat, flat illustration" -o boat.png --aspect square
  431. ```
  432. `--transparent`, `--format png|jpeg|webp` and `--ref FILE` (a reference image,
  433. repeatable) are optional. One image takes about 30 seconds. The subscription
  434. backend picks the resolution and quality itself, so the shape is asked for in
  435. words (`--aspect`). The command is a client of `POST /api/image` on the admin
  436. port, which answers with the image bytes. Generated workers are told the `curl`
  437. form of it, so a worker with only Bash can make an image and read it back.
  438. Claude Code runs `WebSearch` as a separate small-model request. To keep a
  439. ChatGPT-routed setup from spending Anthropic quota for that request, select the
  440. same ChatGPT provider as the search backend in `config.json`:
  441. ```jsonc
  442. { "webSearch": { "provider": "chatgpt", "model": "gpt-5.6-terra" } }
  443. ```
  444. This is opt-in. With no `webSearch` setting, Claude Code's existing search path
  445. is unchanged.
  446. When a slot is routed to a provider that cannot run server tools (an
  447. `openai-compatible` one, for example) and no search backend is set, its searches
  448. are refused rather than answered with invented results. To let only those
  449. searches go to Anthropic while the slot's other requests stay routed, set:
  450. ```jsonc
  451. { "webSearch": "native" }
  452. ```
  453. The search is then sent unchanged, with its original Claude model id, on the
  454. app's own login. A search for a model id that does not start with `claude-` is
  455. still refused. `webSearch` takes either a backend object or `"native"`, not both.
  456. ### Remote (SSH) sessions
  457. Claude Desktop's Code tab can run a session on a remote Linux host over SSH.
  458. Claude Code then runs **on that host**, where ClaudeRipple is not, so routed
  459. models are unknown there (`Agent type 'gpt-6-1-sol' not found`). Connect the
  460. host once from the machine ClaudeRipple runs on:
  461. ```bash
  462. clauderipple remote add my-server # an ssh alias or user@host; --port N (default 18791)
  463. clauderipple remote hosts # tunnel state per host (also in `remote status`)
  464. clauderipple remote remove my-server # undo it
  465. ```
  466. `remote add` needs key-based `ssh my-server` to work without a password prompt.
  467. It copies the router's CA certificate (the public certificate only) to
  468. `~/.clauderipple-remote/` on the host, points the host's
  469. `~/.claude/settings.json` at the tunnel (`HTTPS_PROXY=http://127.0.0.1:18791`,
  470. `NODE_EXTRA_CA_CERTS`; the original is kept as
  471. `settings.json.clauderipple-backup`; a different proxy already there is refused
  472. without `--force`), and copies the generated agent files. The router then keeps
  473. a reverse SSH tunnel open from the host's `127.0.0.1:18791` to itself, restarts
  474. it when it drops, and copies the agent files again whenever they change. The
  475. host needs only `sh` — no Node and no ClaudeRipple.
  476. While the tunnel is down, Claude Code on the host gets *connection refused*
  477. rather than going to Anthropic around the router. The remote session only exists
  478. while this machine's Claude Desktop is connected to it anyway. The tunnel
  479. listens on the host's loopback only, but on a shared host every user there can
  480. reach that port, and through it your router. The status mod and image
  481. generation are not available on remote hosts yet.
  482. ## Lid closed and the phone hotspot
  483. For sessions you drive from your phone (Remote Control, Ripple Remote): the laptop
  484. keeps working in a bag. Open the dashboard's **Laptop** view.
  485. **Stay awake with the lid closed.** Three modes: *Off*; *While working* — awake while a
  486. model call runs and for 10 minutes after the last one; *Always* — until you turn it off
  487. or a time you set. It lets go by itself when the battery drops below 20% or (on macOS) the
  488. Mac reports a thermal warning, and puts a closed laptop to sleep when it does. Don't carry
  489. it in a closed bag for long while it is awake.
  490. - macOS: press **Set up** once. macOS's own password dialog installs a sudoers rule for
  491. exactly the commands used (`pmset disablesleep`, `pmset sleepnow`, joining the hotspot
  492. you name); ClaudeRipple never sees the password. `clauderipple awake setup --remove`
  493. takes it out. If Insomnia is installed, setup removes it (it does the same job).
  494. - Windows: no administrator rights. A child of the router holds
  495. `SetThreadExecutionState`, and the lid action is set to "do nothing" only while it is
  496. held; your own power settings come back afterwards, also if the router dies.
  497. **Wi-Fi without internet → phone hotspot.** It probes the internet through the Wi-Fi
  498. interface (a router with a dead uplink or a captive portal still looks "connected" to
  499. the OS) and joins the hotspot you name after 3 of 5 rounds fail. On the hotspot it tries
  500. the way back 5, 10, 20, 30, 40, 50 and 60 minutes after landing, then every 10 minutes:
  501. it leaves the hotspot so the OS joins the best known Wi-Fi in range, and stays there only
  502. if that Wi-Fi has internet. A model call that starts during the switch waits a moment
  503. instead of failing.
  504. - Join the phone's hotspot from the laptop once by hand, so the OS keeps its password.
  505. - iPhone: keep Settings › Personal Hotspot › "Allow Others to Join" on. On a Mac, the
  506. same Apple Account and Bluetooth on help it find the phone.
  507. - Android: an idle hotspot turns itself off after a while; turn it on again if it is
  508. not found.
  509. - Linux: not supported yet.
  510. ### Codex app and Codex CLI
  511. ```bash
  512. clauderipple codex on # adds a "clauderipple" provider to ~/.codex/config.toml (backup first)
  513. codex --profile clauderipple -m claude-sonnet-5
  514. ```
  515. `codex on` also points Codex's built-in OpenAI provider at ClaudeRipple (`openai_base_url`).
  516. Codex keeps its ChatGPT sign-in; its GPT requests go on to ChatGPT unchanged, except that the
  517. account is chosen from the ones added to ClaudeRipple — so with several added, Codex too moves
  518. from a spent account to the next. With none added, or all of them at their limit, Codex's own
  519. login is used. An `openai_base_url` you set yourself is left alone. `clauderipple codex off` undoes it.
  520. Claude models show up in Codex's model list under their names (ClaudeRipple writes
  521. a model catalog next to Codex's own). Claude is reached through your Claude Code
  522. login (detected from the running Desktop session, the terminal login, or a
  523. sign-in of ClaudeRipple's own: **Providers → Claude (Anthropic) → + Add
  524. Claude account** opens the browser, no terminal needed; the same as
  525. `clauderipple claude-login`) or an Anthropic API key. Reusing a
  526. subscription login is subject to Anthropic's terms. Any Anthropic-compatible
  527. provider you configured is available the same way.
  528. <p align="center">
  529. <img src="docs/media/add-provider.png" width="880" alt="Add provider: ChatGPT subscription, presets, OpenAI-compatible providers">
  530. </p>
  531. ## Providers
  532. | Provider | Kind | Auth | Model list | Notes |
  533. |---|---|---|---|---|
  534. | ChatGPT subscription | Codex backend | several sign-ins, switching when one runs out (Codex login reused too) | read from your subscription (GPT-6 Sol, Luna, Astra, GPT-5.6 …) | effort low…max (Luna: ultra), prompt cache 94–99 % |
  535. | Google Gemini | Gemini API | AI Studio API key, or Google account sign-in (Antigravity, see the FAQ) | discovered | translated (Messages ⇄ generateContent), thought signatures kept across turns |
  536. | Grok subscription | Grok CLI chat proxy | the Grok CLI's sign-in (`grok login`), refreshed through the CLI | read from your subscription (Grok 4.7, 4.7 Fast, 4.6, 4.5) | translated (Chat Completions), effort low…xhigh; the CLI's own channel, not a published API ([§4e](docs/ARCHITECTURE.md#4e-grok-providers-implemented-2026-10-08)) |
  537. | OpenRouter | Anthropic-compatible | API key | 400+, discovered | per-model effort support read from the API |
  538. | DeepSeek, Kimi, Z.ai GLM, MiniMax, Qwen (intl / cn) | Anthropic-compatible | API key | preset | verified against vendor docs |
  539. | xAI Grok, Mistral, Groq, Together, Fireworks | OpenAI-compatible | API key | discovered | translated (Chat Completions / Responses) |
  540. | Ollama, LM Studio | OpenAI-compatible, local | none | discovered | |
  541. | Anthropic | native | Claude login(s) or API key | Claude models | optional sticky rotation for Claude Desktop/Code; Codex picks one login |
  542. | Anything else | custom | your choice | discovered | any Anthropic- or OpenAI-compatible endpoint |
  543. Requests to compatible providers are cleaned of Anthropic-only fields
  544. (server-side threads, deferred tools, context management, thinking bindings) and
  545. effort is clamped per model, so vendors do not 400 on Claude Code's request shape.
  546. ## How it works
  547. ```
  548. Claude Desktop / claude CLI ──HTTPS_PROXY──▶ ClaudeRipple ──▶ api.anthropic.com (unchanged)
  549. │
  550. mapped model ──────────┼──▶ chatgpt.com/backend-api (Responses ⇄ Messages)
  551. ├──▶ Anthropic-compatible vendors (+ compat layer)
  552. ├──▶ OpenAI-compatible vendors (Messages ⇄ Chat/Responses)
  553. └──▶ Google Gemini (Messages ⇄ generateContent)
  554. Codex app / CLI ──/v1/responses──▶ ClaudeRipple ingress ──▶ Claude (your login or API key) / vendors
  555. ```
  556. - The Claude Code CLI reads `HTTPS_PROXY` and `NODE_EXTRA_CA_CERTS` from
  557. `~/.claude/settings.json` (Anthropic's documented corporate-proxy path). Only that
  558. process trusts ClaudeRipple's local CA; the OS keychain is untouched unless you
  559. turn on picker mode.
  560. - Picker mode routes the app's own claude.ai traffic through the proxy and adds
  561. your models to the picker list the app fetches at start. Off again with one click.
  562. - The router drains in-flight calls on restart, self-exits on repeated upstream
  563. failures so launchd restarts it, rotates logs, and never drops a request silently.
  564. Details with sources: [docs/ARCHITECTURE.md](docs/ARCHITECTURE.md).
  565. ## Privacy
  566. Everything runs on 127.0.0.1. API keys live in `~/.clauderipple/config.json`
  567. and added Claude OAuth grants in `~/.clauderipple/claude-accounts.json` (both
  568. 0600). Admin APIs and logs expose neither tokens nor upstream account ids. The only
  569. network destinations are the providers you configure. There is no telemetry.
  570. ## Status: alpha
  571. In daily use by the author, but young. Expect rough edges:
  572. - **Windows support is new (2026-09-14).** Install, setup, the router, signing
  573. in, the picker and a live GPT call are verified on x64 hardware; crash recovery
  574. on arm64. The window still has the stock Windows title bar. Windows builds are
  575. unsigned; see the note under Install.
  576. - **Linux support is new (2026-09-24)**, contributed with measurements on Ubuntu
  577. 24.04 x64 (install, a Code-tab request, restart mid-stream, picker mode, uninstall).
  578. It needs a systemd user session; other distributions and arm64 are untested.
  579. - ChatGPT, OpenRouter and Claude-in-Codex are verified with live accounts; the
  580. other presets follow the vendors' official documentation.
  581. - A model added to the picker is usable from the next session.
  582. - Claude Code and Codex change their wire protocols often; a client update can
  583. break a translation until ClaudeRipple catches up. Bugs and logs welcome.
  584. ## Community
  585. Questions and bug reports are on Discord: [https://discord.gg/R8ZgGSNHjw](https://discord.gg/R8ZgGSNHjw).
  586. What is planned next: [docs/ROADMAP.md](docs/ROADMAP.md).
  587. ## Not affiliated
  588. ClaudeRipple is an independent open-source project. It is not affiliated with,
  589. endorsed by, or sponsored by Anthropic or OpenAI. Claude and Claude Code are
  590. trademarks of Anthropic, PBC. ChatGPT and Codex are trademarks of OpenAI.
  591. ## License
  592. Copyright (c) 2026 pbj. GPL-3.0 — see [LICENSE](LICENSE). Use it freely; if you distribute a modified
  593. version, ship its source under the same license.

Ratings and comments

Sign in to rate this skill.

Sign in

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