Skip to content
 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

282 Commits
 
 
 
 
 
 

Repository files navigation

AdaL CLI


About

AdaL is an AI coding agent that runs in your terminal, created by SylphAI and named after Ada Lovelace, the world's first programmer. It works with Claude, GPT, Gemini, GLM, Kimi, DeepSeek, MiniMax, and local models.

AdaL CLI is a terminal-native platform for orchestrating AI agents, models, and tools into autonomous, end-to-end workflows with persistent memory and human alignment.

This repository is the community home for AdaL CLI — report issues, share what you build, and contribute skills.

Install AdaL CLI

Native install is recommended: it manages AdaL's runtime and updates consistently across platforms.

macOS, Linux, WSL:

curl -fsSL https://adal.sylph.ai/install.sh | bash

Windows PowerShell:

irm https://adal.sylph.ai/install/windows | iex

Windows CMD:

powershell -NoProfile -ExecutionPolicy Bypass -Command "irm https://adal.sylph.ai/install/windows | iex"

If irm is not recognized, you are in CMD rather than PowerShell — use the Windows CMD command above, or open PowerShell.

Then go to your working directory and run:

adal

First run opens the browser for authentication. See the Quickstart for details.

The CLI

Start where developers move fastest. AdaL CLI brings agentic work into the terminal — with model switching, reviewable tool use, session memory, and IDE handoff.

  • Terminal-native — command discovery, tool confirmations, streaming work, and clean recovery paths.
  • Review-first — tool calls, file edits, diffs, and plans stay visible, so you understand the work before signing off.
  • Model freedom — switch between Claude, Gemini, GLM, Kimi, DeepSeek, MiniMax and more from a fast command palette.
  • CLI → IDE handoff — open the same session in AdaL's agentic IDE when a task needs files, terminal, and canvas together.
Command What it does
/model Choose the best model for the job
/ide Open the current session in the agentic IDE
/resume Return to previous work with context intact
/stats Inspect session health, model, and usage

More on the CLI product page.

Documentation

Everything is published at docs.sylph.ai.

Get started

Page What it covers
Quickstart Install, authenticate, first run
Models & Billing Supported models and pricing
Input Methods Files, images, pasted content
Workflows & Examples Real end-to-end usage

Features

Page What it covers
Slash Commands · Keyboard Shortcuts Drive the CLI fast
Skills & Plugins Extend the agent
MCP Servers Model Context Protocol
Deep Research · Web Search Research from the terminal
Browser Use Drive a real browser
Image Generation & Analysis Create and read images
Headless Mode · Cron Scripting, CI, schedules
Local Models · BYOAK Run local or bring your own keys
Manage & Resume Sessions History, stats, resuming work

Also: Changelog · Troubleshooting

Blog

Let's have a journey on agents and models — blog.sylph.ai.

Post
Loop Engineering For Everyone Making loop engineering simple and affordable enough for every engineer
True Autonomous Coding Agents Require Browser Use Real production bugs live in cookies, hydration, redirects, and clicks
AdaL Is 100% Aligned with Andrej Karpathy Human understanding is the bottleneck of agentic engineering
The Ultimate Guide to Agentic Tool Calling What really happens between the model and the JSON tool call
The Last Harness You'll Ever Build The hard part is no longer the model — it is everything around it

📝 Contributing

Skills are the best way to contribute. A skill is a markdown file that teaches the agent a repeatable workflow — no need to touch the CLI internals.

Skills live in SylphAI-Inc/skills (MIT) and install straight into the CLI:

/plugin marketplace add SylphAI-Inc/skills
/plugin install core-skills@adal-agent-skills
/skills   # see what you have

Contribute a Skill

  1. Fork SylphAI-Inc/skills
  2. Create your skill at skills/<your-skill-name>/SKILL.md
  3. Register it in .claude-plugin/marketplace.json
  4. Open a pull request
---
name: skill-name
description: Brief description for the skills list
author: your-username
version: 1.0.0
---

# Skill Title

## When to Use

Describe the trigger conditions.

## Instructions

Step-by-step guidance for the agent.

New to it? Install create-skill and let AdaL scaffold one with you. For a larger example, see swe-cli-skills.

Share Your Use Case

The best way to share a workflow is to ship it as a skill — then anyone can install it in one command. If you have a process that works, package it up and send it over.

Open a pull request →

In the PR description, tell us:

  • What you were trying to do
  • How you used AdaL — prompts, skills, models, commands
  • What worked, and what didn't
  • Anything you would want next

Screenshots and terminal recordings welcome. Prefer to chat first? Post in Discord or tag @adalagent on X.

🐛 Reporting Issues

Open an Issue →

Please include:

  • What you expected vs what happened
  • Steps to reproduce (if applicable)
  • AdaL CLI version (adal -v)
  • OS and terminal

Links


Built by AdaL & SylphAI

About

The self-evolving coding agent that learns from your entire team and codebase. Less syncing. Less waiting. Deliver at the speed of thought.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors