FileCat

Overview#

FileCat is a web-based comprehensive server management tool built on Node.js (TypeScript/JavaScript), open-sourced under Apache-2.0. Latest version: v5.35.2 (798+ commits, 39 releases).

Traditional server ops rely on fragmented tools like WinSCP, terminal clients, and monitoring dashboards with high learning curves. FileCat provides a unified web interface covering file operations, terminal interaction, AI-driven command execution, Docker/process monitoring, multi-protocol proxying, and lightweight CI/CD.

Core Capabilities#

File Management & Collaboration#

• Chunked resumable upload with multiple root directories
• Built-in code editor and image editor modes
• Whiteboard drawing and file sharing

Terminal & Security#

• Web remote terminal based on xterm.js
• Virtual Shell layer with command permission filtering to block dangerous commands like rm -r /

AI Agent#

• Natural language interaction to execute server commands and retrieve system resource info
• AI-executed commands protected by the same permission filtering mechanism
• Compatible with OpenAI-style APIs
• Local knowledge base full-text search (RAG) to enhance AI context

Multi-Protocol Network Proxy & Networking#

• SSH/HTTP proxy for managing multiple Linux servers (similar to WinSCP), Windows RDP remote desktop proxy
• Client/server virtual TUN network for cross-public/private network LAN formation with virtual IPs
• Server-controlled TCP communication proxy through clients

System Monitoring & Automation#

• Docker container management and process monitoring (optimized for low CPU usage)
• Sharded file reading for instant opening of large log files
• Built-in CI/CD using .act files mimicking GitHub workflow syntax
• Multi-user permission system and website navigation bookmarks

Architecture Highlights#

• Base framework: Node.js monolithic architecture with xterm.js terminal component on the frontend
• Terminal security: xterm.js handles UI rendering; backend virtual Shell intercepts and filters commands via user-configured permission whitelists
• AI integration: Pluggable OpenAI-compatible API with local RAG full-text search, injecting retrieval results into AI prompts
• Networking: Built-in TUN virtual NIC networking (client/server) and control-channel-based TCP proxy forwarding
• Large file handling: Sharded reading strategy to avoid memory pressure and frontend rendering lag
• Design inspiration: filebrowser (file management), MeshCentral (remote control), mstsc (RDP)

Installation#

Requires Node.js 20+. If GitHub access is slow, Python 3.8 and a C++ compiler are needed; Windows requires Python 3 + VS Build Tools.

NPM (recommended)

npm install -g filecat

Linux one-liner

curl -o install.sh https://filecat.xiaobaidadada.fun/files/linux-install.sh && bash install.sh

Docker

docker run -d --name filecat --restart=always --net=host -v /home:/home ghcr.io/xiaobaidadada/filecat:latest --port 5567 --base_folder /home

Run & upgrade

filecat --port 5567
filecat-upgrade    # auto-detect environment upgrade from v5.33.0+

Default credentials: admin/admin

Supports filecat --install for one-click systemd service registration (pm2 compatible). Terminal and AI features require pre-configured command permissions and OpenAI-compatible API in user settings.

Unconfirmed Details#

• Specific AI models tested with the Agent are not publicly listed
• RAG vector database selection is unspecified
• TUN network underlying implementation (wireguard/custom) is undisclosed
• No explicit SLA or commercial support terms