Skip to main content

Command Palette

Search for a command to run...

How to Set Up Claude Code CLI with Ace Data Cloud (Beginner's Guide)

A practical terminal-first guide to configuring Claude Code with an API base URL and project-level workflows.

Updated
6 min readView as Markdown
How to Set Up Claude Code CLI with Ace Data Cloud (Beginner's Guide)

If you like working from the terminal, Claude Code becomes much more useful when it is configured once and then available inside every project; this guide walks through a simple setup using Ace Data Cloud as the API base URL.

Claude Code CLI with Ace Data Cloud

What we are setting up

Claude Code is an agentic coding tool that you can run with the claude command from a normal terminal. The useful part is not just “chatting with a model.” It is that the tool can sit inside a real project directory, read context, help explain code, review diffs, write tests, and turn a short instruction into a concrete development task.

In this setup, Ace Data Cloud is used as the API base URL for Claude Code. The two key environment variables from the integration guide are:

export ANTHROPIC_AUTH_TOKEN="{token}"
export ANTHROPIC_BASE_URL="https://api.acedata.cloud"

ANTHROPIC_AUTH_TOKEN is the token you copy from the Ace Data Cloud console. ANTHROPIC_BASE_URL points Claude Code to https://api.acedata.cloud. The guide also notes that the token value is sent with a Bearer prefix automatically, so you do not need to add that prefix yourself in the variable.

This is the kind of setup I like as a builder: small surface area, easy to reason about, and no special editor workflow required. If your terminal works, your coding assistant works.

Install Claude Code

The source guide lists three installation paths. The native installer is the recommended route and works across macOS, Linux, and WSL:

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

On Windows PowerShell, the equivalent command is:

irm https://claude.ai/install.ps1 | iex

And on Windows CMD:

curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd

There are also package-manager options if you prefer them:

brew install claude-code

or, on Windows:

winget install Claude.ClaudeCode

After installation, open a new terminal. If claude is not found, the usual first checks are to restart the terminal, inspect your PATH, or rerun the installer.

Configure the API variables

The most straightforward configuration is to add the Ace Data Cloud variables to your shell profile. For zsh users, that usually means ~/.zshrc; for bash users, it may be ~/.bashrc or ~/.bash_profile.

# Ace Data Cloud - Claude Code Proxy Configuration
export ANTHROPIC_AUTH_TOKEN="{token}"
export ANTHROPIC_BASE_URL="https://api.acedata.cloud"

Then reload your shell configuration:

source ~/.zshrc  # or source ~/.bashrc

If you prefer to keep these variables scoped only to Claude Code, the guide also supports putting them into ~/.claude/settings.json:

{
  "env": {
    "ANTHROPIC_AUTH_TOKEN": "{token}",
    "ANTHROPIC_BASE_URL": "https://api.acedata.cloud"
  }
}

I usually choose the settings.json approach when I do not want global environment variables leaking into unrelated scripts. For a dedicated dev machine, the shell profile method is simpler.

Start inside a real project

Once the variables are set, move into a project directory and start Claude Code:

cd /path/to/your/project
claude

That opens interactive mode. From there, you can ask project-level questions or assign practical tasks, for example:

what does this project do?
there's a bug where users can submit empty forms - fix it
write unit tests for the calculator functions

Claude Code also supports one-off commands directly from the shell:

claude "fix build error"
claude -p "explain this function"
claude -c
claude -r
claude commit

Those commands are useful when you want a quick answer or action without staying in an interactive session. claude -c continues the most recent conversation in the current directory, while claude -r restores a previous conversation.

Inside interactive mode, a few built-in commands are worth remembering:

  • /help shows help information.
  • /clear clears conversation history.
  • /config opens the settings panel.
  • /model switches model.
  • /mcp manages MCP services.
  • /compact compresses context.
  • /memory manages memory.
  • /login switches account.

That /mcp command is especially useful if you are building more automated workflows around local tools, repos, or internal systems.

Use it in scripts and reviews

The terminal version is not limited to interactive chat. It can read piped input, which makes it fit naturally into existing developer habits.

For example, you can review a branch diff:

git diff main | claude -p "review these changes"

Or ask it to watch log output and respond to anomalies:

tail -f app.log | claude -p "notify me if an anomaly is detected"

For teams, I would start with the review workflow. It is narrow enough to be safe, but still saves real time: you can ask for risky changes, missing tests, confusing naming, or migration concerns before opening a pull request.

Add project memory with CLAUDE.md

A small CLAUDE.md file in the project root can make responses much more consistent. Claude Code loads this file at startup, so it is a good place for project-specific expectations:

# Project Description
This is a full-stack project using Django + Vue.js.

## Coding Standards
- Use Python 3.12
- Follow PEP 8 coding style
- All APIs need to have unit tests

This does not need to be long. In practice, a short file with stack details, test commands, naming rules, and “things not to change casually” is enough to prevent many avoidable back-and-forth turns.

Troubleshooting checklist

If Claude Code does not connect after setup, check the basics first:

  1. Confirm your environment variables are set correctly.
  2. Restart the terminal so the shell profile is reloaded.
  3. If needed, create or verify ~/.claude/config.json with {"primaryApiKey": "self"}.
  4. Confirm your Ace Data Cloud token is valid in the console.

The main idea is simple: keep Claude Code close to where development already happens, and keep the configuration explicit. That gives you a practical terminal assistant for explaining code, reviewing diffs, writing tests, and handling small refactors without turning your workflow into a new platform to manage.

If you want the original reference while setting it up, the full Ace Data Cloud document is here: Claude Code Terminal CLI Integration Guide.

More from this blog

A

Ace Data Cloud Blog — AI Infrastructure, 200+ Models, One API

164 posts

The official blog of Ace Data Cloud, one API key and one balance for every major AI model. Chat, image, video, music, and search, all pay-per-use and priced below going direct. Autonomous agents can even pay per call through native x402. Here we share product launches, deep dives into how our infrastructure works, real usage numbers, and practical guides to building on Ace. Written to be clear and useful, whether you're a developer shipping fast, a builder, or just AI-curious.