Agent Station vs ZyCin · AI SSH Remote Coding

Agent Station is an SSH terminal client built for the AI era. Connect to your own servers, manage tmux sessions, and drive AI coding agents like Claude Code and Codex with your keyboard or voice — from your commute, a cafe, anywhere. Key features: • Real SSH terminal: modern xterm.js-based UI with large, high-contrast text; multi-hop ProxyJump support reaches hosts behind bastions • AI agent command deck: a project list shows each agent's live status (working / waiting / disconnected); tap in for a live terminal • Voice input: hold to talk, mixed Chinese/English recognition with optional LLM correction, transcripts injected straight into the remote terminal • Persistent tmux sessions: network drops, app switches, device changes — your session survives • Private by design: SSH credentials never leave your device, zero data collection; optional SSH-over-443 tunneling for restricted networks Who it's for: • Engineers running AI agents like Claude Code or Codex on remote servers • Developers who need to check on long-running server tasks from anywhere • Anyone who wants a clean, modern, ad-free SSH client Requirements: you need your own SSH server. Voice input requires your own speech-recognition API key (Volcengine Doubao ASR supported).

Code from your phone. Connect to your server. On the road, on the train, or putting out a fire — no laptop needed. Tell ZyCin what you want in plain language. It connects to your server over SSH, reads the code, edits it, runs commands and tests, and gets the job done. CORE FEATURES · Conversation as development Describe a task in one sentence. The agent runs the full loop — read files → analyze → edit → verify — with every tool call and code change visible in real time. · Direct SSH to your server Password or private key login. Works with Linux, macOS and Windows (PowerShell). Your code stays on your server — never routed through us. · Multi-protocol file manager Browse and edit over SSH / SFTP / FTP / FTPS / WebDAV, with an icon grid, syntax highlighting and in-place saving. · Switch models freely 175+ built-in provider presets (OpenAI, DeepSeek, Anthropic, Gemini, Kimi and more). Switch models mid-conversation. · Scan for local LLMs Auto-discover Ollama / LM Studio / llama.cpp / vLLM on your LAN — free, offline, and your data never leaves the network. · Image and voice input Send a screenshot so the AI can see the error or the UI. Hold to talk and assign tasks by voice. · MCP extensions Connect MCP (Model Context Protocol) servers to give the agent databases, APIs, Kubernetes and more. · Secure by design Passwords, keys and API keys are stored only in the device secure storage. Writes and command execution ask for confirmation by default, with dangerous commands highlighted. Host fingerprints are pinned on first connect and flagged if they change. WHO IT IS FOR · DevOps and backend engineers who must react to production issues anywhere · Developers who review code and fix bugs on the commute · Indie developers and remote teams NOTES · Bring your own model API key, or use a local model on your LAN. Connecting to a server requires you to have the appropriate access. · New users get a 30-day free trial; afterwards the free tier allows 40 messages per day. Terms of Use (EULA): https://www.apple.com/legal/internet-services/itunes/dev/stdeula/ Privacy Policy: https://www.2cyai.com/privacy.html