Overview

NexusCLI is a lightweight AI cockpit that orchestrates Claude Code, Codex CLI, Gemini CLI, and Qwen Code from a single web and terminal interface.

The project is runtime-aware:

• each engine can expose native and custom lanes
• model selection is tied to a concrete runtime
• the UI can inspect runtime availability and update state
• sessions and messages persist runtime metadata in the local database

NexusCLI is npm-first and targets Linux, macOS, and Termux without requiring desktop-native builds for the core application.

Features

• Multi-engine orchestration for Claude, Codex, Gemini, and Qwen
• Runtime-aware model catalog with native and custom lanes
• SSE streaming with realtime tool and status updates
• Interrupt and resume support per engine
• Session import and history sync from native CLI stores
• Workspace isolation, switching, and history
• File and image attachments where supported
• Runtime inventory API and UI runtime manager
• Conversation search, bookmark/pin, and job runner API
• Voice input support with HTTPS auto-setup

Current Model Support

| Engine | Lane | Examples | |--------|------|----------| | Claude | Native | sonnet, opus, haiku | | Claude | Custom | deepseek-*, glm-4.7, glm-5, qwen3.5-plus, qwen3-max-2026-01-23, kimi-k2.5, MiniMax-M2.7 | | Codex | Native | gpt-5.4, gpt-5.3-codex, gpt-5.2-codex, gpt-5.1-codex-max, codex-mini-latest | | Codex | Custom | qwen3-coder-plus, qwen3-coder-next, qwen3.5-plus, glm-5, deepseek-ai/DeepSeek-V3.2-TEE | | Gemini | Native | gemini-3-pro-preview, gemini-3-flash-preview, gemini-2.5-pro, gemini-2.5-flash | | Qwen | Native | qwen3-coder-plus, qwen3-coder-next, qwen3.5-plus, qwen3-max | | Qwen | Custom | glm-4.7, kimi-k2.5 |

See docs/RUNTIME_MODEL.md for the runtime model and provider mapping.

Install

npm install -g @mmmbuto/nexuscli

Then initialize:

nexuscli init

And start the server:

nexuscli start

Runtime Management

NexusCLI separates:

• engine
• lane
• runtime
• provider
• model

Runtime inventory is available through:

• UI runtime manager
• GET /api/v1/runtimes
• POST /api/v1/runtimes/check
• POST /api/v1/runtimes/install
• POST /api/v1/runtimes/update

API Keys

Provider keys can be stored locally via:

nexuscli api list
nexuscli api set deepseek <key>
nexuscli api set zai <key>
nexuscli api set alibaba <key>
nexuscli api set chutes <key>
nexuscli api set minimax <key>
nexuscli api set openai <key>

These keys are used only for custom provider lanes that need compatible API routing.

Commands

| Command | Description | |---------|-------------| | nexuscli init | Setup wizard | | nexuscli start | Start server | | nexuscli stop | Stop server | | nexuscli status | Show server status | | nexuscli engines | Inspect/configure runtime-aware engines | | nexuscli model | Set/get default model | | nexuscli config | Read/edit configuration | | nexuscli api | Manage provider API keys | | nexuscli workspaces | Manage workspaces | | nexuscli logs | View server logs | | nexuscli setup-termux | Termux bootstrap helpers | | nexuscli update | Update NexusCLI | | nexuscli uninstall | Remove NexusCLI |

Network Access

| Protocol | Default Port | Use Case | |----------|--------------|----------| | HTTP | 41800 | Local access | | HTTPS | 41801 | Remote access and browser microphone support |

HTTPS certificates are auto-generated on first run.

Development

git clone <upstream-or-fork-url> nexuscli
cd nexuscli
npm install
cd frontend && npm install && npm run build && cd ..
npm run dev

Documentation

• docs/ARCHITECTURE.md
• docs/API.md
• docs/GUIDE.md
• docs/RUNTIME_MODEL.md
• docs/PTY_ARCHITECTURE.md

License

MIT License. See LICENSE.