MCP Directory
ServersClientsBlog

xASO - App Store Optimization

AI-powered App Store Optimization platform for mobile apps

Go to xASO
MCP Directory

Model Context Protocol Directory

MKSF LTD
Suite 8805 5 Brayford Square
London, E1 0SG

MCP Directory

  • About
  • Blog
  • Documentation
  • Contact

Menu

  • Servers
  • Clients

© 2026 model-context-protocol.com

The Model Context Protocol (MCP) is an open standard for AI model communication.
Powered by Mert KoseogluSoftware Forge
  1. Home
  2. Clients
  3. CodePilot

CodePilot

GitHub
Website

A multi-model AI agent desktop client — connect any AI provider, extend with MCP & skills, control from your phone. Built with Electron + Next.js.

6,076
644

<img src="docs/icon-readme.png" width="32" height="32" alt="CodePilot" style="vertical-align: middle; margin-right: 8px;" /> CodePilot

A multi-model AI agent desktop client -- connect any AI provider, extend with MCP & skills, control from your phone, and let your assistant learn your workflow.

GitHub release
Downloads
GitHub stars
Platform
License

中文文档 | 日本語


CodePilot


Download | Quick Start | Documentation | Contributing | Community


Download

PlatformDownloadArchitecture
macOSApple Silicon (.dmg) · Intel (.dmg)arm64 / x64
WindowsInstaller (.exe)x64 + arm64
LinuxBuild from sourcex64 + arm64

Or visit the Releases page for all versions.


Why CodePilot

Multi-provider, one interface

Connect to 17+ AI providers out of the box. Switch providers and models mid-conversation without losing context.

CategoryProviders
Direct APIAnthropic, OpenRouter
Cloud platformsAWS Bedrock, Google Vertex AI
Chinese AI providersZhipu GLM (CN/Global), Kimi, Moonshot, MiniMax (CN/Global), Volcengine Ark (Doubao), Xiaomi MiMo, Aliyun Bailian (Qwen)
Local & self-hostedOllama, LiteLLM
CustomAny Anthropic-compatible or OpenAI-compatible endpoint
MediaGoogle Gemini (image generation)

Beyond coding — a full AI agent

CodePilot started as a coding tool but has grown into a general-purpose AI agent desktop:

  • Assistant Workspace — Persona files, persistent memory, onboarding flows, and daily check-ins. Your assistant learns your preferences and adapts over time.
  • Generative UI — AI can create interactive dashboards, charts, and visual widgets rendered live in-app.
  • Remote Bridge — Connect to Telegram, Feishu, Discord, QQ, and WeChat. Send messages from your phone, get responses on your desktop.
  • MCP + Skills — Add MCP servers (stdio / sse / http) with runtime monitoring. Define reusable skills or install from the skills.sh marketplace.
  • Media Studio — AI image generation with batch tasks, gallery, and tagging.
  • Task Scheduler — Schedule recurring tasks with cron expressions or intervals.

Built for daily use

  • Pause, resume, and rewind sessions to any checkpoint
  • Split-screen dual sessions side by side
  • Track token usage and costs with daily charts
  • Import Claude Code CLI session history
  • Dark / Light theme toggle
  • English + Chinese interface

Quick Start

Path A: Download a release (most users)

  1. Download the installer for your platform from the Download section above
  2. Launch CodePilot
  3. Configure a Provider in Settings > Providers — add your API key for any supported provider
  4. Start a conversation

Note: Installing the Claude Code CLI (npm install -g @anthropic-ai/claude-code) unlocks additional capabilities like direct file editing, terminal commands, and git operations. It is recommended but not required for basic chat.

Path B: Build from source (developers)

PrerequisiteMinimum version
Node.js18+
npm9+ (ships with Node 18)
git clone https://github.com/op7418/CodePilot.git
cd CodePilot
npm install
npm run dev              # browser mode at http://localhost:3000
# -- or --
npm run electron:dev     # full desktop app

Core Capabilities

Conversation & Interaction

CapabilityDetails
Interaction modesCode / Plan / Ask
Reasoning effortLow / Medium / High / Max + Thinking mode
Permission controlDefault / Full Access, per-action approval
Session controlPause, resume, rewind to checkpoint, archive
Model switchingChange model mid-conversation
Split screenSide-by-side dual sessions
AttachmentsFiles and images with multimodal vision support
Slash commands/help /clear /cost /compact /doctor /review and more

Extensions & Integrations

CapabilityDetails
Providers17+ providers: Anthropic, OpenRouter, Bedrock, Vertex, Zhipu GLM, Kimi, Moonshot, MiniMax, Volcengine, MiMo, Bailian, Ollama, LiteLLM, custom endpoints
MCP serversstdio / sse / http, runtime status monitoring
SkillsCustom / project / global skills, skills.sh marketplace
BridgeTelegram / Feishu / Discord / QQ / WeChat remote control
CLI importImport Claude Code CLI .jsonl session history
Image generationGemini image gen, batch tasks, gallery

Data & Workspace

CapabilityDetails
Assistant WorkspacePersona files (soul.md, user.md, claude.md, memory.md), onboarding, daily check-ins, persistent memory
Generative UIAI-created interactive dashboards and visual widgets
File browserProject file tree with syntax-highlighted preview
Git panelStatus, branches, commits, worktree management
Usage analyticsToken counts, cost estimates, daily usage charts
Task schedulerCron-based and interval scheduling with persistence
Local storageSQLite (WAL mode), all data stays on your machine
i18nEnglish + Chinese
ThemesDark / Light, one-click toggle

First Launch

  1. Configure a Provider — Go to Settings > Providers and add credentials for the provider you want to use. CodePilot includes presets for all major providers — just pick one and enter your API key.
  2. Create a conversation — Pick a working directory, select a mode (Code / Plan / Ask), and choose a model.
  3. Set up Assistant Workspace (optional) — Go to Settings > Assistant, choose a workspace directory, and enable Onboarding. CodePilot creates soul.md, user.md, claude.md, and memory.md at the workspace root.
  4. Add MCP servers (optional) — Go to the MCP page in the sidebar to add and manage MCP servers. Custom skills are managed on the separate Skills page.
  5. Install Claude Code CLI (optional) — For advanced features like file editing and terminal commands, install the CLI: npm install -g @anthropic-ai/claude-code

Platform & Installation Notes

macOS builds are code-signed with a Developer ID certificate but not notarized, so Gatekeeper may still prompt on first launch. Windows and Linux builds are unsigned.

<details> <summary>macOS: Gatekeeper warning on first launch</summary>

Option 1 -- Right-click CodePilot.app in Finder > Open > confirm.

Option 2 -- System Settings > Privacy & Security > scroll to Security > click Open Anyway.

Option 3 -- Run in Terminal:

xattr -cr /Applications/CodePilot.app
</details> <details> <summary>Windows: SmartScreen blocks the installer</summary>

Option 1 -- Click "More info" on the SmartScreen dialog, then "Run anyway".

Option 2 -- Settings > Apps > Advanced app settings > set App Install Control to allow apps from anywhere.

</details>

Documentation

📖 Full documentation: English | 中文

Getting started:

  • Quick Start -- Download or build from source
  • First Launch -- Provider setup, workspace configuration
  • Installation Guide -- Detailed setup instructions

User guides:

  • Providers -- Configuring AI providers and custom endpoints
  • MCP Servers -- Adding and managing Model Context Protocol servers
  • Skills -- Custom skills, project skills, and the skills.sh marketplace
  • Bridge -- Remote control via Telegram, Feishu, Discord, QQ, WeChat
  • Assistant Workspace -- Persona files, onboarding, memory, daily check-ins
  • FAQ -- Common issues and solutions

Developer docs:

  • ARCHITECTURE.md -- Architecture, tech stack, directory structure, data flow
  • docs/handover/ -- Design decisions and handover documents
  • docs/exec-plans/ -- Execution plans and tech debt tracker

FAQ

<details> <summary>Do I need the Claude Code CLI?</summary>

No. You can use CodePilot with any supported provider (OpenRouter, Zhipu GLM, Volcengine, Ollama, etc.) without the Claude Code CLI. The CLI is only needed if you want Claude to directly edit files, run terminal commands, or use git operations on your machine. For chat and assistant features, just configure a provider and start a conversation.

</details> <details> <summary>Configured a Provider but no models appear</summary>

Verify the API key is valid and the endpoint is reachable. Some providers (Bedrock, Vertex) require additional environment variables or IAM configuration beyond the API key. Use the built-in diagnostics (Settings > Providers > Run Diagnostics) to check connectivity.

</details> <details> <summary>What is the difference between <code>npm run dev</code> and <code>npm run electron:dev</code>?</summary>

npm run dev starts only the Next.js dev server -- you use CodePilot in your browser at http://localhost:3000. npm run electron:dev starts both Next.js and the Electron shell, giving you the full desktop app experience with native window controls.

</details> <details> <summary>Where are the Assistant Workspace files?</summary>

When you set up a workspace, CodePilot creates four Markdown files at the workspace root directory: soul.md (personality), user.md (user profile), claude.md (rules), and memory.md (long-term notes). State tracking (onboarding progress, check-in dates) is stored in the .assistant/ subdirectory. Daily memories go to memory/daily/.

</details> <details> <summary>Bridge requires additional setup per platform</summary>

Each Bridge channel (Telegram, Feishu, Discord, QQ, WeChat) requires its own bot token or app credentials. Go to the Bridge page in the sidebar to configure channels. You will need to create a bot on the target platform first and provide the token to CodePilot.

</details>

Community

<img src="docs/wechat-group-qr.png" width="240" alt="WeChat Group QR Code" />

Scan the QR code to join the WeChat user group for discussions, feedback, and updates.

  • GitHub Issues -- Bug reports and feature requests
  • GitHub Discussions -- Questions and general discussion

Contributing

  1. Fork the repository and create a feature branch
  2. npm install and npm run electron:dev to develop locally
  3. Run npm run test before opening a PR
  4. Submit a PR against main with a clear description

Keep PRs focused -- one feature or fix per pull request.

<details> <summary>Development commands</summary>
npm run dev                    # Next.js dev server (browser)
npm run electron:dev           # Full Electron app (dev mode)
npm run build                  # Production build
npm run electron:build         # Build Electron distributable
npm run electron:pack:mac      # macOS DMG (arm64 + x64)
npm run electron:pack:win      # Windows NSIS installer
npm run electron:pack:linux    # Linux AppImage, deb, rpm (local build only — not a GitHub Release download)

CI/CD: Pushing a v* tag builds the macOS and Windows packages and creates a GitHub Release automatically. Linux is not part of the release build — build it locally with npm run electron:pack:linux.

Notes:

  • Electron forks a Next.js standalone server on 127.0.0.1 with a random free port
  • Chat data is stored in ~/.codepilot/codepilot.db (dev mode: ./data/)
  • SQLite uses WAL mode for fast concurrent reads</details>

License

Business Source License 1.1 (BSL-1.1)

  • Personal / academic / non-profit use: free and unrestricted
  • Commercial use: requires a separate license — contact @op7418 on X
  • Change date: 2029-03-16 — after which the code converts to Apache 2.0

Repository

OP
op7418

op7418/CodePilot

Created

February 6, 2026

Updated

June 15, 2026

Language

TypeScript

Category

AI