Setup Guide

New Ora Setup Guide

Install your API access in minutes using the setup command from your order delivery message. Works with Claude Code CLI and Codex CLI-compatible tools.

OpenAI-compatible endpoint
This is not an official OpenAI, Anthropic, or OpenLimits API key/account. New Ora provides OpenAI-compatible API access through a custom endpoint. Upstream provider keys are never shared with customers.

Before you start

  • You have an active New Ora order.
  • You have the setup command sent after your order is prepared.
  • You will run the command in your terminal (PowerShell / Terminal).
  • After setup, run Claude Code CLI from your project folder.
Do not share your access
Do not share your setup command or installed access. If the same access is detected on more than the allowed user/device limit, API access may be suspended.

Windows PowerShell setup

One-line install using the setup token from your delivery message.

powershell
irm "https://newora.space/api/v1/setup?token=YOUR_API_KEY" | iex
  1. Open Windows PowerShell.
  2. Paste the setup command from your delivery message.
  3. Press Enter and wait for New Ora Setup to finish.
  4. Confirm the setup output shows Customer token installed, OPENAI_API_KEY / OPENAI_BASE_URL configured, Package active, Connection test successful, Usage tracking enabled.
  5. Open your project folder and run claude.
If PowerShell blocks execution
Open PowerShell as a normal user first (no admin needed). Paste the one-line command exactly as provided. Do not manually edit the token. If it still fails, contact support with your order number.

macOS / Linux setup

Terminal one-liner using the setup token from your delivery message.

bash
curl -fsSL "https://newora.space/api/v1/setup?token=YOUR_API_KEY" | sh
  1. Open Terminal.
  2. Paste the macOS/Linux setup command from your delivery message.
  3. Press Enter and wait for setup to finish.
  4. Restart your terminal if needed.
  5. Open your project folder and run claude.

Manual API setup

For tools that need OpenAI-compatible provider configuration.

API Type
OpenAI-compatible
Base URL
https://newora.space/ol/v1
API Key
Installed by your setup command, or view it in your customer dashboard when enabled.
Use only the New Ora endpoint
Do not use the upstream OpenLimits/OpenAI URL. Use only the New Ora endpoint above.

Claude Code CLI usage

After setup, launch Claude Code from your project folder.

  1. Go to your project directory.
  2. Run claude.
  3. Send a quick test message such as hello.
  4. If the assistant replies, setup is working correctly.

Codex CLI usage

Codex-compatible tools use standard OpenAI environment variables.

Configure your Codex-compatible tool with:

  • OPENAI_API_KEY — installed by your setup command
  • OPENAI_BASE_URLhttps://newora.space/ol/v1

Or configure a manual OpenAI-compatible provider using the same Base URL.

How to verify setup

Expected output from a successful setup run.

output
New Ora Setup
Unlimited Access Gateway

Gateway:   https://api.newora.space/v1
Customer:  Customer #0042
Order:     NO-123456789
Plan:      Economy Daily
Status:    Active
Access:    High Usage / Fair Usage
Expires:   2026-07-04
Token:     newora_live...A92F

[OK] Customer token installed
[OK] OPENAI_API_KEY configured
[OK] OPENAI_BASE_URL configured
[OK] Package active
[OK] Connection test successful
[OK] Usage tracking enabled

Done! Run 'claude' in any project directory.

Common setup problems

Most issues fall into one of these five buckets.

"Invalid API key"
Possible causes
  • setup token expired
  • order access expired
  • copied wrong command
  • customer access disabled/suspended
Solution

Regenerate setup from support/admin. Contact support with your order number.

"Connection test failed"
Possible causes
  • internet issue
  • package expired
  • temporary provider issue
  • wrong terminal command
Solution

Retry after a few minutes. Verify the command. Contact support if it persists.

"claude command not found"
Possible causes
  • Claude Code CLI is not installed
Solution

Install Claude Code CLI separately. New Ora setup only configures API environment variables.

"OPENAI_BASE_URL not detected"
Possible causes
  • terminal not restarted after setup
Solution

Restart your terminal. Rerun the setup command. Verify environment variables.

"Access suspended"
Possible causes
  • suspected sharing or abuse
Solution

Open a support ticket with your order number. Support remains available for suspended users.

Quota, model families & error codes

How the daily USD allowance is enforced and what to expect when it runs out.

Model families
  • fable_5 — Fable 5 family (any *fable* alias plus the exact upstream model id where supported).
  • opus_4_8 — Opus 4.8 family (any claude-opus-4* alias plus the exact upstream model id where supported).
  • Unsupported models return 400 unsupported_model.
Quota behavior

Each active API key has a rolling 24-hour window per model family. USD-equivalent cost is calculated from token usage and deducted from your daily allowance. When you hit the allowance for that family, further calls to that family return 429 quota_exceeded. Other families remain callable if their allowance is still available. Windows reset 24 hours after key activation and every 24 hours after that.

"No fixed daily cap"

Some plans have no per-day USD quota for a given model. Fair-use, abuse protection, security limits, and availability controls still apply. This is not unrestricted access.

Error codes
  • 401 unauthorized — API key missing or invalid.
  • 403 access_expired — package expired.
  • 403 access_suspended — key suspended.
  • 400 unsupported_model — model not mapped to any family.
  • 429 quota_exceeded — daily USD allowance reached for that model family.
  • 503 pool_unavailable — no active upstream in the global managed pool.
  • 502 provider_error — upstream provider returned an error.
Streaming & compatibility

OpenAI-compatible chat/completions endpoints support streaming (stream: true). Both streamed and non-streamed calls are billed the same and count against the same daily USD allowance.

Security

Your API key is treated as a secret — never commit it, share it, or paste it in support chats. Upstream provider keys are managed server-side and never exposed to customers.

Still stuck?

Open a support ticket with your order number and we'll take it from there.

Contact support