Alma Client Setup Guide

A powerful AI coding client.

Overview

Alma is a powerful and polished AI client that integrates code writing, terminal operations, Git management, browser features, and more, giving developers a complete AI-assisted programming experience.

Links

Resource	URL
Alma website	https://alma.now/

Features

Alma provides:

• Code writing: intelligent code generation and completion
• Terminal operations: integrated terminal for direct command execution
• Git management: visual Git operations
• Browser integration: built-in browser support
• Multi-model support: Claude, Gemini, and more
• Custom providers: connect custom API providers

Installation and Configuration

Step 1: Install Alma

Open the Alma website and download the version for your operating system.

Step 2: Set the Interface Language

For a better experience, you can set the interface to Chinese if needed:

1. Open Alma
2. Go to Settings -> General -> Language
3. Choose Chinese

Step 3: Add Xiaomai API as a Custom Provider

1. Open provider settings

Click Select Provider -> Add Custom Provider.

2. Configure the Xiaomai API provider

Fill in:

• Provider Name: Xiaomai API, or any custom name
• Base URL: https://xiaomai.win/v1
• API Key: paste the API key created in the Xiaomai API console

3. Fetch available models

After adding the provider, click Fetch in the Xiaomai API provider screen. Alma retrieves all available models automatically.

Choose the model you need and start using it.

Step 4: Configure Multiple Groups (Optional)

Xiaomai API supports using different API keys, from different groups, to access different model sets.

In the provider screen, you can:

1. Enter API keys from different groups
2. Fetch the model list for each group
3. Switch based on your task

Tip

To use Gemini 3 image-generation capabilities, choose an API key from the default channel.

New Chats and Usage

Important Context Setting

When creating a new chat, the context range setting is important.

Strong Recommendation

Choose All instead of Auto.

Testing showed that Auto can cause serious hallucinations, such as:

• Referencing files that do not exist
• Generating incorrect code logic
• Misunderstanding project structure

Choosing All gives the model complete context and improves accuracy.

Usage Tips

1. Code writing

• Describe your requirement and Alma generates code
• Supports multi-file and project-level changes
• Generated code can be previewed and edited

2. Terminal operations

• Run commands in the integrated terminal
• View command output
• Debug issues with AI assistance

3. Git management

• View Git status visually
• Let AI help write commit messages
• Manage branches and merges

4. File browsing

• Browse project files
• Locate and edit files quickly
• Search and filter files

Best Practices

1. Choose the Right Model

• Claude Sonnet: daily coding tasks, faster responses
• Claude Opus: complex architecture and hard problems
• Gemini: multimodal tasks, such as image understanding

2. Use Context Carefully

• For small projects, select all files
• For large projects, select relevant modules and files
• Clear irrelevant context regularly to avoid token waste

3. Use Alma's Integrated Features

• Use the terminal to verify AI-generated code
• Use Git integration for version control
• Use the browser to view docs and references

FAQ

How do I switch API providers?

Use the model selector at the top of the chat screen to switch providers and models anytime.

Why are some models unavailable?

• Check whether the API key has permission for the model
• Confirm account balance is sufficient
• Check token-group settings in the Xiaomai API console. Claude uses groups starting with C, while Gemini uses the gemini group

What is the difference between Alma and Claude Code?

Feature	Alma	Claude Code
Interface	Graphical UI	Command-line UI
Best for	Full-featured development environment	Fast terminal operations
Integrated features	Terminal, Git, browser, and more	Pure CLI tool
Learning curve	Lower	Higher

Choose based on your workflow.

More Questions

Read the FAQ or contact support.

Done

Configuration is complete. Remember:

• Select All for context range
• Choose models based on task
• Use Alma's integrated features
• Check API balance regularly