Usage Guide
AI Coding Tools Installation & Configuration Guide
Platform API Base URL
v1OPENAI · COMPATIBLE · ENDPOINT
$https://your-domain.com/api/nexura/v1select · tools
Choose your tool
Only Claude Code is available now. More tools are coming soon.
install · script
Installation script
Run this script in your terminal
curl -s https://nexura.startupper.space/setup-claudecode.sh | bash -s -- --url https://nexura.startupper.space/api/nexura --key nxr-xxxxxxxxxxxxThis script automatically detects your system, installs required dependencies, and configures environment variables. After it finishes, please reopen the terminal to start using the tool.
Run this script in PowerShell as Administrator
& { $url="https://nexura.startupper.space/api/nexura"; $key="nxr-xxxxxxxxxxxx"; iex ([Text.Encoding]::UTF8.GetString((New-Object Net.WebClient).DownloadData('https://nexura.startupper.space/setup-claudecode.ps1'))) }This script automatically detects your system, installs required dependencies, and configures environment variables. After it finishes, please reopen PowerShell. If you see a permission error, run [Set-ExecutionPolicy RemoteSigned -Scope CurrentUser -Force] as Administrator and try again.
catalog · models
Available Models
Below are the currently available models on the platform. Use the Model ID when configuring your tools.
reference · errors
Error Code Reference
| Error Code | HTTP | Meaning | Common Cause | Solution |
|---|---|---|---|---|
AI0001 | 400 | Invalid Request | Malformed request body or missing required fields | Check JSON format and required parameters |
AI0002 | 401 | Invalid API Key | API Key does not exist or has wrong format | Verify API Key and environment variable settings |
AI0003 | 403 | API Key Disabled | Key has been disabled or revoked by admin | Contact admin or create a new API Key |
AI0004 | 403 | Quota Exhausted | Insufficient available quota | Purchase a package or top up quota |
AI0005 | 429 | Rate Limited | Exceeded API Key or group RPM limit | Reduce request frequency or upgrade group |
AI0006 | 400 | Model Not Found | Requested model ID not in platform model list | Check model name spelling, refer to the model list below |
AI0007 | 403 | Model Not Allowed | Current group or package does not include this model | Upgrade package or contact admin for access |
AI0008 | 502 | Upstream Error | Provider API returned an error | Retry later; contact admin if persistent |
AI0009 | 504 | Upstream Timeout | Provider API response timed out | Reduce input length or retry later |
AI0010 | 503 | No Available Channel | All channels are unavailable or under maintenance | Retry later; contact admin if persistent |
AI0011 | 400 | Parameter Validation Failed | Request parameters do not meet requirements | Check parameter types and value ranges |
AI0012 | 413 | Content Too Large | Request body exceeds size limit | Reduce input content length |
AI0013 | 500 | Internal Server Error | Platform internal exception | Retry later; contact admin if persistent |
AI0014 | 402 | Insufficient Balance | Balance insufficient during quota pre-deduction | Top up or purchase a package |
AI0015 | 429 | Concurrent Limit Exceeded | Too many simultaneous requests for your group | Wait for current requests to complete before sending new ones |
help · faq
FAQ
What if the 'claude' command is not found?
Windows users: make sure PowerShell was run as Administrator during installation. macOS/Linux users: reopen the terminal or run source ~/.zshrc (or ~/.bashrc). If installed via npm, ensure the Node.js global bin directory is in your PATH.
What if the 'codex' command is not found?
Verify Node.js version ≥ 18 (run node --version). Ensure the npm global install directory is in your system PATH. Windows users can try restarting PowerShell. macOS/Linux users can run npm config get prefix to check the global directory.
What if the tool fails to connect to the platform?
Check: 1) OPENAI_API_KEY and OPENAI_BASE_URL environment variables are set correctly; 2) The API Key is valid (not revoked); 3) Your network can reach the platform URL. You can test connectivity with a curl command.
How to specify which model to use in the tool?
Claude Code uses the --model flag (e.g., claude --provider openai --model gpt-4o). Codex CLI uses the --model flag (e.g., codex --model gpt-4o). See the model list above for available models.
What if I get an insufficient quota error during use?
Check your remaining quota on the Dashboard. If exhausted, purchase a package or use a redemption code. AI coding tools consume many tokens — choose a package with sufficient quota.
How to handle rate limiting (429) errors?
Tools usually retry automatically. If it happens frequently, your request rate exceeds the group limit. Suggestions: 1) Reduce concurrent operations; 2) Upgrade to a higher group for higher RPM limits.