relaykite
Open-source ChatGPT & Codex agents for Android and Windows. AI phone control, app automation, files and PowerShell via a private MCP relay. Android works without an always-on PC.
Install / Use
claude mcp add Stervakim -- npx -y github:Stervakim/relaykiteIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
AutomationSupported Platforms
Tags
Our assessment of relaykite
relaykite scores 78/100 on our quality scale, 2552nd of 2,857 Automation skills we index.
Its MCP Server is 8.3 KB long, well organised into 9 sections with 1 code example: a thorough specification that gives an agent plenty to work with.
It has 3 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated today, so relaykite is actively maintained.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 92/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-11-01. Automated pattern scan on 2026-11-01. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
relaykite compared with similar skills
All 4 of these similar skills score higher than relaykite; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| relaykite (this skill)by Stervakim | 78 | 3 | today | MCP Server |
| Agent-Reachby Panniantong | 100 | 107.3k | 7d ago | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 78.5k | 1d ago | CLAUDE.md |
| CowAgentby zhayujie | 100 | 47.7k | today | CLAUDE.md |
| Scraplingby D4Vinci | 100 | 91.2k | today | MCP Server |
Frequently asked questions
- How do I install relaykite?
- Run
claude mcp add Stervakim -- npx -y github:Stervakim/relaykite. The install tabs above show the steps for each supported agent. - Which AI agents does relaykite work with?
- It is written for Claude Code, Claude Desktop and OpenAI Codex, as a MCP Server file. Other agents that read the same format can often use it too.
- Is relaykite safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 92/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is relaykite still maintained?
- The repository was last updated today, so relaykite is actively maintained.
Skill content
View source on GitHubRelayKite — AI tasks. Your devices.

Let ChatGPT tap, swipe and run tasks on your devices.
Control your Android phone and Windows PC through your own private MCP relay.
Website · Download v0.3.0 · Android setup · Windows setup · Русский
RelayKite gives your AI assistant a way to act on the devices you own. Ask it to inspect your phone, open an app, work with files, or run an authorized PowerShell task on your PC. ChatGPT plans the task; a local agent carries it out and returns the result.
The Android agent runs on the phone itself. After setup, everyday phone control does not require a permanently running computer, USB or desktop ADB. The Windows agent runs independently on its own PC. Both need internet and your private relay.
Previously Codaki Mini Codex. Version 0.3.0 adds autonomous original Codex to the existing device agent. Existing ZIP names, module IDs, tool names, credentials and installation paths retain their legacy names for compatibility.
What is it for?
Use RelayKite when you want a conversation to lead to an action on your own device:
| You ask | The device agent can do | | --- | --- | | “Check my phone's battery and available storage.” | Return device status and inspect storage through available tools. | | “Open Settings and tell me what is on the screen.” | Launch the app, read the UI or capture a screenshot. | | “Save this note in my chosen folder.” | Write a file within the configured access. | | “Diagnose why my phone agent stopped.” | Inspect relevant services and reuse saved maintenance notes. | | “List large files in this Windows folder.” | Run an authorized PowerShell task on the connected PC. |
These are example requests, not recorded demos. Actual results depend on your device, configured permissions and available tools.
Choose your device
| | Android | Windows |
| --- | --- | --- |
| Local worker | APK-embedded Bionic Python/Node | Python background agent + PowerShell |
| Requirements | Android app; existing Magisk for root modes; Termux optional | Python 3.14+, PowerShell and selected permissions |
| Connection | Your authenticated HTTPS MCP relay and supported ChatGPT plugin | Your authenticated HTTPS MCP relay and supported ChatGPT plugin |
| Does it need the other device? | No permanently running PC after setup | No phone required |
| Install | Signed APK; matching Magisk ZIP for root modes | mini-codex-pc-0.3.0.zip |
| Guide | Android installation | Windows installation |
How it works
You → ChatGPT → your personal plugin → your private HTTPS relay
↕
Local device agent
Android or Windows
↓
Action → result → chat
The model runs in ChatGPT. Ordinary device tools return their result to the conversation; autonomous development adds an on-device Codex supervisor that requests further model decisions through a saved conversation after the outer response ends. RelayKite does not host model weights or remove account limits. Memory stores notes and verified solutions between tasks; it does not train model weights. Architecture and private relay setup.
Get started
- Check the requirements. Autonomous Android development needs existing root/Magisk, the signed runtime and the Portable component; Termux is optional; Windows needs Python 3.14+ and PowerShell. Check that your ChatGPT account supports the required personal plugin connection.
- Download the right installer and
SHA256SUMSfrom release v0.3.0. Check the SHA-256 checksum before installation. - Follow the guide for your device. Configure your own private relay and credentials, then connect your personal plugin. Importing a public card alone does not connect your device.
- Verify with a small task. Start with device status, inspect a test app or list a test folder, and check the actual result.
For assisted installation, give Codex the release ZIP and checksum file:
Install RelayKite v0.3.0 from this ZIP on my chosen device. Read INSTALL_WITH_CODEX.md, verify SHA256SUMS and requirements, configure my own private relay, and verify the connection. Use the device tools available to you. Preserve existing credentials, pairings, permissions, memory and logs during an update.
Initial Android preparation may need a computer. The installer does not install root or enable unavailable account features. Compatibility · Access and data.
Source status after 0.3.0 (not yet a packaged release)
The main source now pins verified upstream Codex CLI 0.165.0 for the next release.
This was checked against the official GitHub release archive and its extracted binary SHA-256,
and the ARM64 Linux-musl executable passed --version and app-server --help
on the tested rooted Android phone. The installed author's 0.165.0 binary is an
Android-target build with different bytes, so executable identity or complete
end-to-end parity is not claimed. The published v0.3.0 ZIP remains on
Codex 0.154.0 until new signed and tested packages are published. Do not mix
ZIP/APK components from unrelated versions or assume a GitHub source commit
automatically updates your installed agent.
Release 0.3.0
Original Codex 0.154.0 runs on the phone with native patches, shell, Git and semantic Android MCP tools. A durable supervisor continues a task after the outer ChatGPT answer, records checkpoints and receipts, and supports pause, resume and cancellation. Model requests use a normal saved ChatGPT conversation; no API key or temporary chat is required for this transport. The installed ChatGPT app, an unlocked device, network access and account capacity remain necessary.
Context is bounded and can use original Codex compaction. Compatible Codex skills and configured MCP servers can be used; ChatGPT plugins do not automatically become Codex tools. Android root execution is unsandboxed. Native Android model image input is disabled in this release. Architecture and operation · Verification and limits.
Changelog · Boot and proxy guide · Optional memory runtime · Portable session instructions
Learn, verify and contribute
- Control Android from ChatGPT without a running PC — how the architecture works and how to check it.
- Windows guide — installation, grants and independent operation.
- Android without a PC: FAQ.
- Fact sheet for reviewers — source-backed claims and a reproducible test plan.
- Contributing · Report a bug · Brand assets · English launch copy and headline kit.
Version 0.3.0 was exercised on Redmi Turbo 3 / Android 17 / Magisk using embedded Python/Node, without a Termux runtime. Host/mock checks complement real Codex patch, shell, Git, compaction and Android MCP acceptance; other devices still require testing. Android virtual-display support depends on separately installed components. Do not assume offline AI, lock-screen bypass or identical support on every device.
Codaki Android Use and Codaki Portable Use are separate companion repositories.
License and support
RelayKite is an independent open-source project under the MIT license. It is not an official OpenAI product. Source and public releases are free; your ChatGPT account and relay hosting may have their own costs. Support development on Boosty.
Related Skills
Agent-Reach
107.3kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
headroom
78.5kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
CowAgent
47.7kOpen-source personal AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-agent, multi-model, multi-channel. Lightweight, extensible, one-line install.
Scrapling
91.2k🕷️ The BEST Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ and follow here for daily tips and tricks: https://x.com/Scrapling_dev
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
