ailoha.
Home/ Docs/ AI-assisted setup

AI-assisted setup

Let your coding agent do the install, wire-up, and verification. Three commands, one prompt.

Tools Copilot · Claude Code · Cursor · Codex · Cline Time ~2 minutes Short link ailoha.dev/llms.txt

Why this exists

Adding the agent by hand is straightforward, but if you already work with an AI coding agent it's easier to let it install the package, wire up the debug-only registration, and verify that the agent is running. Ailoha ships small framework-specific skill files plus a single llms.txt at the site root so any LLM-aware tool can pick up the right instructions without you copy-pasting docs.

The three steps

1. Install the CLI

The CLI runs on your dev machine. It handles agent discovery, exposes the MCP server, and ships the framework skills.

See Installing the Ailoha CLI for the curl/iex one-liner and channel options.

Confirm with:

ailoha --version

2. Initialize skills in your project

From your app's repo root, run:

cd path/to/your/app
ailoha init

This drops Markdown skill files into .claude/skills/ by default. Your AI agent reads them automatically the next time it runs in this workspace. Pass --target github, --target agent, or --target agents to write into .github/skills/, .agent/skills/, or .agents/skills/ instead.

i
The CLI auto-detects which platforms (MAUI, Android, Flutter, React Native, Expo, WinUI) are present in the workspace and only installs the relevant skills. Override with --platforms maui,android,flutter,react-native,expo,winui.

3. Prompt your agent

Open Copilot, Claude Code, Cursor, Codex CLI, Cline, or whichever agent you use, and send any of these:

Set up the Ailoha agent in this project.
Add Ailoha to my app and verify the agent comes up in debug builds.
Wire Ailoha into the MAUI head project, debug-only.

The agent will read the skill files installed in step 2, pick the matching framework guide, install the package, register the agent behind the right debug flag, and confirm with ailoha agent status. Review the diff before you commit.

One-shot path: llms.txt

If you don't want to run ailoha init first, or you're working in a tool that doesn't read project-local skill files, point your agent at the hosted ailoha.dev/llms.txt file. It follows the llms.txt convention and contains everything an agent needs to set up Ailoha in a fresh project. 

Read https://ailoha.dev/llms.txt and set up Ailoha in this project.

That single sentence is enough for most agents to do the full install, wire-up, and verification.

★
Even shorter. Once the CLI is installed globally, just say "set up Ailoha in this project" — most coding agents will discover the CLI on PATH, run ailoha init themselves, and proceed.

Verify it worked

After the agent finishes, launch your app in debug mode and run from the project root:

ailoha agent status
ailoha ui tree

agent status returns the agent's metadata and platform info. ui tree dumps the live visual tree of the running app. If both succeed, you're done.

Prefer to do it yourself?

The manual setup guides cover every platform with the same steps your agent would run:

Troubleshooting

  • Agent didn't pick up the skills. Some tools cache the workspace skill list — restart the agent session after running ailoha init. WPF is a package preview and does not have a bundled ailoha init skill yet.
  • Wrong target directory. Re-run with --target matching your tool. claude writes to .claude/skills/, github writes to .github/skills/.
  • CLI not found after install. Make sure ~/.ailoha/bin (or %USERPROFILE%\.ailoha\bin on Windows) is on your PATH.
  • Agent edited release code. Ask it to wrap the registration in #if DEBUG / kDebugMode / __DEV__; the skills tell it to do this but a stricter prompt helps.