Skip to content
andershsueh edited this page Feb 10, 2026 · 1 revision

FAQ - Frequently Asked Questions

Quick answers to common questions about ALICE.

General Questions

What is ALICE?

ALICE is an AI-powered command-line assistant that combines conversational AI with real-world tools. Think of it as GitHub Copilot CLI, but with more capabilities.

What can ALICE do?

  • Answer questions about your codebase
  • Read and search files
  • Execute system commands
  • Get Git repository information
  • Provide coding assistance
  • And more through the Tool System

Is ALICE free?

ALICE itself is open source and free. However, you need an LLM backend:

  • Free: Use LM Studio with local models
  • Paid: Use OpenAI API (pay per use)

What LLMs does ALICE support?

Any OpenAI-compatible API:

  • OpenAI (GPT-3.5, GPT-4)
  • LM Studio (local models)
  • Ollama
  • Any custom OpenAI-compatible endpoint

See Model Configuration →


Installation & Setup

How do I install ALICE?

See the Installation Guide for detailed instructions.

Quick install:

npm install -g alice-cli

Where is the configuration file?

~/.alice/settings.jsonc

Platform-specific paths:

  • Windows: C:\Users\<username>\.alice\settings.jsonc
  • macOS: /Users/<username>/.alice/settings.jsonc
  • Linux: /home/<username>/.alice/settings.jsonc

How do I reset configuration?

Delete the config file:

# macOS/Linux
rm ~/.alice/settings.jsonc

# Windows
del %USERPROFILE%\.alice\settings.jsonc

ALICE will recreate it with defaults on next run.


Using ALICE

How do I start ALICE?

alice

Or from source:

npm run dev

How do I exit ALICE?

Press Ctrl+C or Ctrl+D

Can I run ALICE without a banner?

Yes, use the --no-banner flag:

alice --no-banner

How do I clear the screen?

Type /clear or press Ctrl+L

Does ALICE have command history?

Yes! Use and arrows to navigate history.

Can ALICE remember previous conversations?

Yes, within the current session. Conversation history is lost when you exit.

Session persistence is planned for future releases.


Tool System

What are tools?

Tools are functions that let ALICE interact with your system:

  • Read files
  • Search directories
  • Execute commands
  • Get Git info
  • And more

Learn about the Tool System →

How do I know when a tool is being used?

ALICE shows real-time status:

[🔧 readFile] Reading package.json...
[✅ readFile] Success (1024 bytes)

Can I disable tools?

Not yet. All builtin tools are always enabled.

Tool configuration is planned for future releases.

Can I create custom tools?

Not yet. Plugin system is planned for v0.3.0.

What's a "dangerous command"?

Commands that can cause data loss or system changes:

  • File deletion (rm -rf)
  • Disk formatting
  • System shutdown
  • etc.

Learn about dangerous commands →

How do I disable dangerous command warnings?

Edit ~/.alice/settings.jsonc:

{
  "dangerous_cmd": false
}

⚠️ Warning: Only do this if you know what you're doing!


LLM & API

Do I need an API key?

Depends on your LLM backend:

  • LM Studio: No API key needed ✅
  • OpenAI API: Yes, required ❌
  • Ollama: No API key needed ✅

Where do I put my API key?

In ~/.alice/settings.jsonc:

{
  "llm": {
    "apiKey": "sk-your-key-here"
  }
}

Or use environment variable:

{
  "llm": {
    "apiKey": "$OPENAI_API_KEY"
  }
}

Learn about secure API key management →

How do I change the model?

Edit ~/.alice/settings.jsonc:

{
  "llm": {
    "model": "gpt-4"
  }
}

Restart ALICE for changes to take effect.

Can I use multiple models?

Yes, but you need to edit the config and restart ALICE each time.

Multi-model switching is planned for future releases.

Why are responses slow?

Possible reasons:

  1. Large model - Try a smaller model
  2. High temperature - Lower to 0.3-0.5
  3. Network latency - Check your connection
  4. LM Studio performance - Local models vary in speed

Can I make responses faster?

Yes:

  • Use smaller models (e.g., GPT-3.5 instead of GPT-4)
  • Lower maxTokens in config
  • Lower temperature for more focused responses
  • Use local LM Studio for no network latency

Errors & Troubleshooting

Error: "Failed to connect to LLM"

Causes:

  • LM Studio not running
  • Wrong baseURL in config
  • Network issues

Solutions:

  1. Start LM Studio
  2. Verify baseURL: http://localhost:1234/v1
  3. Check if port 1234 is accessible

See full troubleshooting guide →

Error: "Invalid API key"

Causes:

  • Wrong API key
  • Environment variable not set
  • API key has expired

Solutions:

  1. Verify API key is correct
  2. Check environment variable: echo $OPENAI_API_KEY
  3. Generate new API key from OpenAI dashboard

Error: "Model not found"

Causes:

  • Model not loaded in LM Studio
  • Typo in model name
  • Model not available in your plan

Solutions:

  1. Load model in LM Studio
  2. Check model name spelling
  3. Use "auto" to let LM Studio choose

Error: "Tool execution failed"

Causes:

  • File not found
  • Permission denied
  • Invalid parameters

Solutions:

  1. Check file paths are correct
  2. Verify you have necessary permissions
  3. Look at error message for details

ALICE is not responding

Try:

  1. Press Ctrl+C to interrupt
  2. Restart ALICE
  3. Check LLM backend is running
  4. Check terminal for error messages

Strange characters in output

Causes:

  • Terminal doesn't support Unicode
  • ANSI escape codes not working

Solutions:

  1. Use a modern terminal (Windows Terminal, iTerm2, etc.)
  2. Update terminal settings
  3. Try different terminal emulator

Features

Does ALICE support streaming?

Yes! Responses stream in real-time by default.

Can ALICE edit files?

Not yet. Currently ALICE can only:

  • ✅ Read files
  • ✅ Search files
  • ✅ List directories
  • ❌ Write/edit files (planned)

Does ALICE have a GUI?

No, ALICE is command-line only. That's the point! 😊

Can I use ALICE in scripts?

Not recommended. ALICE is designed for interactive use.

For automation, consider using the LLM API directly.

Does ALICE work offline?

Partially:

  • ✅ Local tools work offline
  • ❌ LLM requires network (unless using local LM Studio)

Can ALICE browse the web?

No. ALICE is focused on local development tasks.

Web search may be added in future versions.


Development

Can I contribute to ALICE?

Yes! ALICE is open source: https://github.com/AndersHsueh/Alice

How do I build from source?

git clone https://github.com/AndersHsueh/Alice.git
cd Alice
npm install
npm run build
npm run dev

See Development Guide →

Where can I report bugs?

File an issue on GitHub: https://github.com/AndersHsueh/Alice/issues

Where can I request features?

Same place - GitHub Issues with "feature request" label.

What's the roadmap?

See the GitHub Issues and Projects.


Comparison

ALICE vs GitHub Copilot CLI

Feature ALICE Copilot CLI
Open Source
Local LLM Support
Tool System ✅ (7 tools) ✅ (limited)
Streaming
Cost Free* Paid

*Free with local LLM, paid with OpenAI

ALICE vs ChatGPT

Feature ALICE ChatGPT
CLI Interface
System Integration
File Access
Command Execution
Web Access
Voice

Performance

How much memory does ALICE use?

Approximately:

  • Base: ~50-100 MB
  • With LLM: Depends on model (local models: 2-8 GB)

Does ALICE slow down my system?

No. ALICE is lightweight. The LLM backend (if local) may use resources.

Can I run ALICE on a Raspberry Pi?

Technically yes, but:

  • ✅ ALICE itself works fine
  • ❌ Local LLMs require powerful hardware
  • ✅ Use cloud LLM API instead

Privacy & Security

Does ALICE send my code to the cloud?

Only if you use cloud-based LLMs (OpenAI, etc.).

With local LM Studio, everything stays on your machine.

Is it safe to use ALICE with proprietary code?

Yes, if using local LLM. Be cautious with cloud LLMs - check your company's policies.

What data does ALICE collect?

None. ALICE doesn't have analytics or telemetry.

Can ALICE access my files without permission?

ALICE only accesses files when:

  1. The LLM decides to use a tool
  2. You can see which tool is being called
  3. You can cancel with Ctrl+C

For dangerous operations, confirmation is required (unless disabled).


Still Have Questions?


Not finding what you need? Let us know and we'll add it to the FAQ!

Clone this wiki locally