🚀 OpsTerm — AI Terminal Assistant

🌍 README.id.md — Baca versi Bahasa Indonesia

OpsTerm is a local AI terminal assistant that lives in your terminal. SSH into any server without losing AI access — because the AI runs on your local terminal, not on the remote server.

Inspired by Warp.dev — but free to use any custom AI provider (DeepSeek, OpenAI, Ollama, OpenRouter, etc.)

✨ Features

Feature	Command	Description
🤖 AI Chat	`opsterm how to check disk`	Ask AI anything
💬 Chat REPL	`opsterm chat`	Interactive chat mode with history
🔑 Smart SSH	`opsterm ssh vps-utama`	SSH without remembering IPs
🔮 SSH Escape	Press `Ctrl+B` during SSH	Ask AI & run server commands without dropping the session
🔗 Connect	`opsterm connect vps-utama`	AI REPL for manual SSH / split-pane users
🔗 Multi-hop SSH	`opsterm ssh internal --via bastion`	SSH through jump host
📁 SCP File Transfer	`opsterm scp file.txt server:/path`	Upload/download via server
⚡ Workflow	`opsterm run deploy-app`	Multi-step automation (SSH/SCP/local)
🔗 Pipe Mode	`docker ps \| opsterm "any errors?"`	Explain command output with AI
🗜️ RTK AI	`auto`	Compress command output 60-95% before AI (saves tokens)
💻 Shell Integration	`opsterm explain-last`	Explain previous command output
🏓 Server Ping	`opsterm servers ping vps`	Check SSH connectivity & latency
🔌 Provider Test	`opsterm provider test openai`	Test AI provider connection
📡 Provider Models	`opsterm provider models openai`	List available models for a provider
⌨️ Tab Completion	`source <(opsterm completion bash)`	Auto-complete servers/workflows
📋 History	`opsterm history`	Command history
🔄 Self-Update	`opsterm update`	Check & install latest version
🛠️ Custom Provider	`opsterm provider add <name> --api-key KEY`	Choose any AI provider
🖥️ Web Dashboard	`opsterm web [--port <port>] [--open]`	Browser UI to manage servers/workflows/config
🏥 Diagnostics	`opsterm doctor`	Check config & diagnose issues
📋 Server Details	`opsterm servers show <name>`	View server connection details
🔄 Server Rename	`opsterm servers rename <old> <new>`	Rename a server
📥 Import SSH Config	`opsterm servers import-ssh-config`	Import servers from ~/.ssh/config
⚙️ Custom System Prompt	`opsterm config set ai.system_prompt <text>`	Customize AI personality
✅ Config Validate	`opsterm config validate`	Validate all YAML config files
📚 Search History	`opsterm search <query>`	Search chat history by keyword
🔄 Chat Resume	`opsterm chat --continue`	Resume last chat session
📦 Config Export	`opsterm export [<file>]`	Export config to tar.gz (keys masked)
📥 Config Import	`opsterm import <file>`	Import config from tar.gz
🗑️ Config Reset	`opsterm reset`	Reset config to defaults
📋 Workflow Init	`opsterm workflows init`	Create sample workflows
🌐 Batch SSH	`opsterm ssh --all <command>`	Run command on all servers

💡 All list commands (provider list, servers list, history) support --json flag for scripting.

🚀 Quick Start

Install

Linux/macOS — single curl command, no repo needed:

# Install latest release
curl -L https://raw.githubusercontent.com/edsuwarna/opsterm/main/LATEST -o /tmp/opsterm_latest
OPSTERM_VER=$(cat /tmp/opsterm_latest)
curl -L "https://raw.githubusercontent.com/edsuwarna/opsterm/${OPSTERM_VER}/bin/opsterm" -o ~/.local/bin/opsterm
chmod +x ~/.local/bin/opsterm
rm -f /tmp/opsterm_latest

Or pin to a specific release version (recommended for stability):

# Install v0.7.0
curl -L https://raw.githubusercontent.com/edsuwarna/opsterm/v0.7.0/bin/opsterm -o ~/.local/bin/opsterm
chmod +x ~/.local/bin/opsterm

Check the Releases page for the latest version.

Make sure ~/.local/bin is in your PATH:

# Add to ~/.bashrc or ~/.zshrc
export PATH="$HOME/.local/bin:$PATH"

Configure your AI provider (one-time)

# Add OpenAI (you can switch to any provider later)
opsterm provider add default --api-key 'sk-...' --model gpt-4o

# See all available providers
opsterm provider supported

# See what you've configured
opsterm provider list

Verify

opsterm --help

Updating

OpsTerm has a built-in self-updater:

# Check for updates & install
opsterm update

This will:

Check the latest version from GitHub
Download and compile-verify the new script
Backup your current version as opsterm.bak
Replace with the new version

💡 OpsTerm also checks for updates automatically once per 24 hours on startup — you'll see a notification if a new version is available.

💡 What Can I Do With It?

🎯 Just ask AI

# Ask for a command
opsterm how to find large files

# Explain system output
opsterm explain-last

# In pipe mode
docker ps | opsterm "any errors?"

🌐 SSH — no more IP juggling

Save your servers once, then SSH by name:

# ~/.opsterm/servers.yaml
servers:
  - name: vps-utama
    host: 203.0.113.10
    user: root
    proxy:                    # optional jump host
      host: 198.51.100.5
      user: ubuntu
  - name: db-server
    host: 10.0.0.5
    user: admin

opsterm ssh vps-utama             # Direct SSH
opsterm ssh internal --via bastion # Via proxy/jump host
opsterm scp file.txt server:/path  # File transfer
opsterm servers ping vps-utama    # Check connectivity

🔮 SSH Escape — AI & commands during SSH

Warp-like SSH experience: press Ctrl+B during an active SSH session to access OpsTerm's AI assistant without disconnecting.

# SSH into a server
opsterm ssh vps-utama

# (inside SSH session) Press Ctrl+B → drops to OpsTerm prompt:

╔══════════════════════════════════════════╗
║  🔮 OpsTerm — AI Escape Mode           ║
║  tanya apa aja, !cmd buat execute,     ║
║  resume/exit buat lanjut/selesai       ║
╚══════════════════════════════════════════╝

opsterm❯ ada berapa container running?
  🔍 Detecting command... running `docker ps -a`
  ✅ Ada 5 container: 3 redis, 2 nginx
opsterm❯ !df -h
  Filesystem      Size  Used Avail Use% Mounted on
  /dev/vda1        20G   12G    8G  60% /
opsterm❯ resume
  🔑 Resuming SSH session...

# → back to interactive SSH

Commands in escape mode:

Input	Behavior
`ada berapa container?`	AI detects if a shell command is needed, runs it, explains the output
`!df -h`	Runs the command directly on the server (bypasses AI)
`resume`	Return to the SSH session
`exit`	Terminate the SSH session

No installation required on the remote server — the second SSH connection (for command execution) is opened from your local machine.

Disable escape mode (falls back to regular execvp SSH):

opsterm config set ssh.escape_key_enabled false

🔗 Connect — AI REPL for manual SSH users

Already SSH'd into a server the old-fashioned way? Open another terminal (or tmux pane) and use opsterm connect to attach AI to that server.

# Terminal 1 — manual SSH
ssh user@vps-utama

# Terminal 2 (or tmux pane) — attach AI
opsterm connect vps-utama

╔══════════════════════════════════════════╗
║  🔮 OpsTerm — Connect: vps-utama       ║
║  root@203.0.113.10                      ║
║  Tanya apa aja jawab pake AI           ║
║  !<cmd> buat execute langsung           ║
║  exit / quit buat selesai               ║
╚══════════════════════════════════════════╝

opsterm❯ ada berapa container running?
  🔍 Running `docker ps -a` via second SSH...
  ✅ Ada 5 container: 3 redis, 2 nginx
opsterm❯ !df -h
  Filesystem      Size  Used Avail Use% Mounted on
  /dev/vda1        20G   12G    8G  60% /
opsterm❯ exit
  👋 Disconnected from vps-utama.

Same AI smarts as the SSH escape mode — no installation needed on the remote server.

⚡ Workflows — automate multi-step tasks

Define reusable workflows in YAML:

# ~/.opsterm/workflows.yaml
workflows:
  - name: deploy-app
    steps:
      - command: "cd /opt/app && git pull && docker compose restart"
      - ssh: vps-utama
        command: "cd /opt/app && docker compose up -d --build"
  - name: check-all
    steps:
      - local: "echo '=== Checking all servers ==='"
      - ssh: vps-utama
        command: "df -h && free -h && uptime"
      - ssh: db-server
        command: "pg_isready && systemctl status postgresql --no-pager"

opsterm run deploy-app
opsterm run check-all

🧠 AI Features

Custom AI Providers

OpsTerm works with any OpenAI-compatible API — bring your own provider:

opsterm provider supported                 # List all supported providers
opsterm provider add openai --api-key sk-... --model gpt-4o
opsterm provider add deepseek --api-key sk-... --model deepseek-chat \
  --api-url https://api.deepseek.com/v1/chat/completions
opsterm provider add ollama --api-url http://localhost:11434/v1  # local (no key needed)

opsterm provider default deepseek           # Switch default
opsterm provider test openai                # Test connection
opsterm provider models openai              # List available models
opsterm provider list                       # See all configured

Auto-compress command output by 60-95% before sending to AI — saves tokens, speeds up responses, lowers cost.

Token Compression via RTK

# Before RTK: pytest output = 597 chars → AI
# After RTK:  pytest output =   18 chars → AI (-96% 🚀)

Feature	Benefit
🔌 Auto-detect	RTK identifies output type (git diff, pytest, docker, logs, etc.)
🎯 Smart compression	Keeps relevant lines, removes noise
⚡ Fast	Sub-second compression, even for large outputs
🟢 Status	Shows `🟢 RTK x.x.x` in `opsterm provider list`

Session Context

OpsTerm maintains conversation context throughout a session (30-min window). Every prompt includes previous exchanges so you can ask follow-ups:

# Ask a follow-up — OpsTerm remembers what you just did!
opsterm how to check disk usage
# → Responds with df -h / du -sh

opsterm what about inodes?
# → Understands you're still talking about disk commands

Custom System Prompt

Set a custom system prompt to tailor the AI's behavior and personality:

# Be concise and technical
opsterm config set ai.system_prompt "You are a senior DevOps engineer. Answer in short, actionable commands."

# Speak in a specific language
opsterm config set ai.system_prompt "Answer in Indonesian. Use casual friendly tone."

# Be educational
opsterm config set ai.system_prompt "Explain concepts like I'm a beginner. Include examples."

💡 The system prompt is prepended to every AI request. Use it to set context, tone, or expertise level.

💻 Usage

opsterm <command> [options]

# Ask for a command
opsterm how to check disk usage

# Generate a command
opsterm generate docker compose restart postgres

Smart SSH

opsterm ssh vps-utama                    # Fuzzy match server name
opsterm ssh vps                          # "vps" matches "vps-utama"
opsterm ssh internal --via bastion       # Multi-hop via ProxyJump
opsterm scp file.txt vps-utama:/home/ubuntu/
opsterm scp vps-utama:/var/log/syslog .
opsterm scp file.txt vps-utama:/tmp --via bastion
opsterm servers ping vps-utama           # Check SSH connectivity
opsterm servers show vps-utama           # Show server details

Workflows

opsterm run <name>
opsterm run deploy-app

Tab Completion

# Temporary (current session only)
source <(opsterm completion bash)   # or: opsterm completion zsh

# Permanent (auto-detect shell, adds to .bashrc/.zshrc)
opsterm completion install

After install, tab-complete server names, workflows, and subcommands: opsterm ssh [Tab] → server names opsterm run [Tab] → workflow names opsterm servers [Tab] → subcommands

Diagnostics

Check your OpsTerm installation and configuration:

opsterm doctor

This checks:

✅ Config file exists and is valid
✅ API key is configured (with masked preview)
✅ Provider URL is reachable
✅ Version status (latest vs installed)
✅ PATH includes ~/.local/bin

Use it when setting up OpsTerm for the first time or troubleshooting issues.

🛠️ Server Management

opsterm servers list           # List all servers (with PROXY column)
opsterm servers show vps       # View server details
opsterm servers add            # Add new server
opsterm servers edit vps       # Edit server
opsterm servers rm vps         # Remove server
opsterm servers ping vps       # Check SSH connectivity & latency

⚙️ Configuration

opsterm config list            # View all config
opsterm config set ai.model deepseek-chat
opsterm config set ai.api_url https://api.deepseek.com/v1/chat/completions
opsterm config set shell.confirm_before_exec true

🔧 Config Files

Config is stored in ~/.opsterm/:

~/.opsterm/
├── config.yaml       # AI provider & shell settings
├── servers.yaml      # Server list (+ proxy jump)
├── workflows.yaml    # Workflow list (SSH/SCP/local)

├── history.db        # History (SQLite, auto)
├── last_output.txt   # Last command output
└── last_command.txt  # Last command

⚡ Changed from ~/.ai-workflows to ~/.opsterm/ — auto-migrates on first run.

Environment variables:

export OPSTERM_DIR="/path/to/custom/config"   # Override config dir
export OPSTERM_API_KEY="sk-..."                # AI API key

🧪 Example Workflows

Basic Deploy

# ~/.opsterm/workflows.yaml
workflows:
  - name: deploy-app
    steps:
      - command: "cd /home/ubuntu/app && docker compose pull && docker compose up -d"
      - command: "echo '✅ Deploy complete!'"

Multi-Server Health Check

workflows:
  - name: health-check
    steps:
      - local: "echo '=== Health Check Report ==='"
      - ssh: web-server
        command: |
          echo "--- Uptime ---"
          uptime
          echo "--- Disk ---"
          df -h /
      - ssh: db-server
        command: "pg_isready && echo 'PostgreSQL OK'"
      - local: "echo '✅ Done!'"

Deploy from CI

workflows:
  - name: deploy-ci
    steps:
      - command: "cd /opt/app && git pull && systemctl restart app"

📋 Roadmap

_{Built with ❤️ for developers who manage too many servers.}
_{Inspired by Warp.dev — AI in your terminal, your own provider.}

License

Apache License 2.0 — see LICENSE.

Name		Name	Last commit message	Last commit date
bin		bin
completions		completions
docs		docs
prd		prd
shell		shell
sketches		sketches
web-template		web-template
zsh		zsh
.gitignore		.gitignore
LATEST		LATEST
LICENSE		LICENSE
README.id.md		README.id.md
README.md		README.md
setup.sh		setup.sh
test-report.html		test-report.html

Navigation Menu

Repository files navigation

🚀 OpsTerm — AI Terminal Assistant

✨ Features

🚀 Quick Start

Install

Configure your AI provider (one-time)

Verify

Updating

💡 What Can I Do With It?

🎯 Just ask AI

🌐 SSH — no more IP juggling

🔮 SSH Escape — AI & commands during SSH

🔗 Connect — AI REPL for manual SSH users

⚡ Workflows — automate multi-step tasks

🧠 AI Features

Custom AI Providers

Token Compression via RTK

Session Context

Custom System Prompt

💻 Usage

Smart SSH

Workflows

Tab Completion

Diagnostics

🛠️ Server Management

⚙️ Configuration

🔧 Config Files

🧪 Example Workflows

Basic Deploy

Multi-Server Health Check

Deploy from CI

📋 Roadmap

License

About

Releases

Packages

Contributors

Languages

Latest commit

History

Folders and files

Repository files navigation

🚀 OpsTerm — AI Terminal Assistant

✨ Features

🚀 Quick Start

Install

Configure your AI provider (one-time)

Verify

Updating

💡 What Can I Do With It?

🎯 Just ask AI

🌐 SSH — no more IP juggling

🔮 SSH Escape — AI & commands during SSH

🔗 Connect — AI REPL for manual SSH users

⚡ Workflows — automate multi-step tasks

🧠 AI Features

Custom AI Providers

Token Compression via RTK

Session Context

Custom System Prompt

💻 Usage

Smart SSH

Workflows

Tab Completion

Diagnostics

🛠️ Server Management

⚙️ Configuration

🔧 Config Files

🧪 Example Workflows

Basic Deploy

Multi-Server Health Check

Deploy from CI

📋 Roadmap

License

About

Resources

Security policy

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages