Auru Key dashboard
AURU ACCOUNT TOOLS

Check your Auru key.

See your remaining balance, recent requests, and the setup details for your tools.

Balance is shown in tokens. Lighter models use less of it, so the same balance lasts longer.

Your API key

Enter the full key, including the sk- prefix.

Saved only in this browser

Models

Use the model names exactly as shown. GPT and Grok models use the OpenAI format; Claude models work with both OpenAI and Anthropic formats.

GPT — chat & coding

gpt-5.6-solgpt-5.6-terragpt-5.6-lunagpt-5.5gpt-5.4

Claude — chat & coding

claude-opus-5claude-opus-4-8claude-opus-4-7claude-opus-4-6claude-sonnet-5claude-sonnet-4-6claude-haiku-4-5-20251001claude-fable-5claude-fable-5-1

Gemini — chat & long context

gemini-2.5-progemini-3.6-flash-tieredgemini-3-flashgemini-2.5-flashgemini-3.1-flash-litegemini-2.5-flash-lite

Grok — chat

grok-4.6grok-4.5

Image generation

gpt-image-2

via /v1/images/generations

Video generation

veo_3_1-fast-landscapeveo_3_1-fast-portrait

via /v1/videos

Setup guides

Choose your client below. Every option uses the same Auru key; only the base URL and environment variable names change.

Base URL
https://auru.cc.cd/v1 (OpenAI) · https://auru.cc.cd (Anthropic SDK)
Auth
your key, starting with sk-
Formats
OpenAI-compatible /v1/chat/completions · Anthropic-compatible /v1/messages
1Quick start · curl

OpenAI format

Works with every model.

curl https://auru.cc.cd/v1/chat/completions \
  -H "Authorization: Bearer $AURU_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "messages": [{"role":"user","content":"Hello!"}]
  }'

Anthropic format

For Claude models and Anthropic-native tools.

curl https://auru.cc.cd/v1/messages \
  -H "x-api-key: $AURU_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-5",
    "max_tokens": 1024,
    "messages": [{"role":"user","content":"Hello!"}]
  }'

Streaming

Add "stream": true to receive server-sent events.

curl https://auru.cc.cd/v1/chat/completions \
  -H "Authorization: Bearer $AURU_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "stream": true,
    "messages": [{"role":"user","content":"Hi"}]
  }'

Image generation

Returns a generated image URL or Base64 data.

curl https://auru.cc.cd/v1/images/generations \
  -H "Authorization: Bearer $AURU_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "a red panda coding at night",
    "size": "1024x1024"
  }'

Video generation

Create a task, poll its returned id, then download the completed video.

curl https://auru.cc.cd/v1/videos \
  -H "Authorization: Bearer $AURU_KEY" \
  -F "model=veo_3_1-fast-landscape" \
  -F "prompt=a paper boat crossing a moonlit lake"

curl https://auru.cc.cd/v1/videos/$VIDEO_ID \
  -H "Authorization: Bearer $AURU_KEY"

curl https://auru.cc.cd/v1/videos/$VIDEO_ID/content \
  -H "Authorization: Bearer $AURU_KEY" \
  -o video.mp4
2SDKs

Python · OpenAI

from openai import OpenAI

client = OpenAI(
    api_key="sk-...",
    base_url="https://auru.cc.cd/v1",
)
r = client.chat.completions.create(
    model="gpt-5.5",
    messages=[{"role":"user","content":"Hi"}],
)
print(r.choices[0].message.content)

Python · Anthropic

from anthropic import Anthropic

client = Anthropic(
    api_key="sk-...",
    base_url="https://auru.cc.cd",
)
msg = client.messages.create(
    model="claude-sonnet-5",
    max_tokens=1024,
    messages=[{"role":"user","content":"Hi"}],
)
print(msg.content[0].text)

Node.js · OpenAI

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.AURU_KEY,
  baseURL: "https://auru.cc.cd/v1",
});
const r = await client.chat.completions.create({
  model: "gpt-5.5",
  messages: [{ role: "user", content: "Hi" }],
});
console.log(r.choices[0].message.content);

Node.js · Anthropic

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  apiKey: process.env.AURU_KEY,
  baseURL: "https://auru.cc.cd",
});
const msg = await client.messages.create({
  model: "claude-sonnet-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hi" }],
});
console.log(msg.content[0].text);

Coding CLIs

Claude CodeCLI · Anthropic formatSupported

Use Claude models through Claude Code’s native Anthropic connection.

  1. Install Claude Code and open a terminal in your project.
  2. Set the Auru endpoint and key for the current terminal session.
export ANTHROPIC_BASE_URL="https://auru.cc.cd"
export ANTHROPIC_AUTH_TOKEN="sk-..."
  1. Run claude to start the client.
  2. Choose a Claude model with /model, using an exact name such as claude-sonnet-5.
  3. Persist the variables if needed by adding them to your private shell profile.
Codex CLICLI · OpenAI formatSupported

Codex reads an OpenAI-compatible endpoint from environment variables.

  1. Install Codex and open the repository you want it to work on.
  2. Set the base URL and key.
export OPENAI_BASE_URL="https://auru.cc.cd/v1"
export OPENAI_API_KEY="sk-..."
  1. Start Codex with codex --model gpt-5.5.
  2. Check Recent usage above after the first request, then grant broader repository permissions if needed.
OpenCodeCLI / TUI · OpenAI formatSupported

Add Auru as an OpenAI-compatible provider in your project config.

  1. Set the key with export AURU_KEY="sk-...".
  2. Create opencode.json in the project.
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "auru": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Auru",
      "options": {
        "baseURL": "https://auru.cc.cd/v1",
        "apiKey": "{env:AURU_KEY}"
      },
      "models": {
        "gpt-5.5": { "name": "GPT 5.5" }
      }
    }
  }
}
  1. Launch OpenCode and select auru/gpt-5.5 from the model picker.
  2. Add more model IDs under models when you need them.
AiderCLI · OpenAI formatSupported

Aider accepts a custom OpenAI base URL directly.

  1. Install Aider and open your repository.
  2. Set the endpoint and key.
export OPENAI_API_BASE="https://auru.cc.cd/v1"
export OPENAI_API_KEY="sk-..."
  1. Start Aider with aider --model openai/gpt-5.5.
  2. Use the exact model ID if Aider reports that the model is unknown.

IDEs and extensions

CursorIDE · OpenAI formatSupported

Cursor’s labels can move between releases, but the required values stay the same.

  1. Open Cursor Settings, then go to Models.
  2. Enable the OpenAI API key option and paste your Auru key.
  3. Turn on Override OpenAI Base URL and enter https://auru.cc.cd/v1.
  4. Add a custom model with an exact ID such as gpt-5.5 or claude-sonnet-5.
  5. Select it in Chat or Agent mode and send a short test prompt.
Cline / Roo CodeVS Code · OpenAI formatSupported

Both extensions expose the same OpenAI-compatible provider fields.

  1. Install Cline or Roo Code and open its VS Code sidebar.
  2. Choose OpenAI Compatible in provider settings.
  3. Set Base URL to https://auru.cc.cd/v1.
  4. Paste your key and enter an exact Model ID.
  5. Test in Ask mode before allowing edits or command execution.
ContinueVS Code / JetBrains · OpenAI formatSupported

Add Auru from Continue’s model configuration screen.

  1. Install Continue and open its sidebar.
  2. Choose Add model, then an OpenAI-compatible provider.
  3. Set API base to https://auru.cc.cd/v1 and paste your key.
  4. Enter a model ID such as gpt-5.5.
  5. Save, select the model, and send a test message.
ZedEditor · OpenAI formatSupported

Zed reads the key from your environment and model details from settings.json.

  1. Set the key before launching Zed with export OPENAI_API_KEY="sk-...".
  2. Open Zed settings.json and add the endpoint.
{
  "language_models": {
    "openai": {
      "api_url": "https://auru.cc.cd/v1",
      "available_models": [{
        "name": "gpt-5.5",
        "display_name": "GPT 5.5",
        "max_tokens": 128000
      }]
    }
  }
}
  1. Restart Zed, open the Agent panel, and select the new model.

Desktop apps

Codex appmacOS · Custom providerSupported

The Codex app shares its provider configuration with the CLI.

  1. Quit Codex so it reloads configuration later.
  2. Store the key for macOS apps with launchctl setenv AURU_KEY "sk-...".
  3. Edit ~/.codex/config.toml.
model = "gpt-5.5"
model_provider = "auru"

[model_providers.auru]
name = "Auru"
base_url = "https://auru.cc.cd/v1"
env_key = "AURU_KEY"
wire_api = "responses"
  1. Reopen Codex and start a small test task.
Cherry StudiomacOS / Windows / Linux · OpenAI formatSupported

Cherry Studio supports custom OpenAI-compatible providers.

  1. Open Settings, then Model Providers.
  2. Add an OpenAI-compatible provider.
  3. Set API Host to https://auru.cc.cd/v1 and paste your key.
  4. Add models manually using exact IDs from the Models section.
  5. Enable the provider and test it in a new chat.
ChatboxDesktop / mobile / web · OpenAI formatSupported

Chatbox can connect to Auru as a custom OpenAI-compatible provider.

  1. Open Settings and find model providers.
  2. Choose a custom OpenAI-compatible provider.
  3. Enter https://auru.cc.cd/v1 as the API host and paste your key.
  4. Enter a model ID from the Models section.
  5. Save and start a new conversation with that model.
Claude DesktopDesktop app · No custom endpointNot available

Claude Desktop does not currently allow a custom Anthropic API base URL. An Auru key cannot be connected directly.

  1. Use Claude Code for Anthropic’s native workflow with Auru.
  2. Use Cherry Studio or Chatbox for desktop chat with Claude models.
  3. Do not put the key in MCP settings; MCP adds tools and does not change Claude Desktop’s model provider.

FAQ

What base URL do I use?

https://auru.cc.cd/v1 for OpenAI-compatible clients, or https://auru.cc.cd for Anthropic SDKs (they append /v1/messages themselves). Authenticate with your sk- key.

Which models can I call?

See the Models section above — five GPT models, nine Claude models, grok-4.5, gpt-image-2, and two Veo 3.1 video models. Call any of them by their exact name. GPT and Grok use the OpenAI format; Claude works in both OpenAI and Anthropic formats.

How is my balance measured?

Your balance uses the reference model's rate. Lighter models deduct fewer tokens, so the same balance lasts longer. Enter your key above to see the exact balance and recent usage.

What happens when I run out of balance?

Requests return an insufficient_user_quota error until you top up. Balance and keys are issued by whoever gave you this key — contact them to add more or to raise a limit.

Do you support streaming, tools, and vision?

Yes — wherever the underlying model supports them. Pass "stream": true, tools/function-calling, or image inputs exactly as you would with OpenAI or Anthropic.

Can I set spend or rate limits on a key?

Per-key request and daily spend caps are set by whoever issued the key. Ask them if you need a limit changed.

My request failed — how do I debug it?

Errors follow the standard OpenAI/Anthropic shape with an HTTP status code and a message. Check the Recent usage table above to confirm the call reached us, verify the model name is spelled exactly, and make sure your key isn't expired or out of balance.

Common ones decoded: a message starting 预扣费额度失败 means pre-charge failed — balance too low for this request (the two numbers are your remaining balance and what the call needed). Invalid token means the key is wrong, disabled, or missing its sk- prefix — paste the full key.

Claude Code says "There's an issue with the selected model"

Claude Code shows that message for any API error, not just model problems. Nine times out of ten it's an empty balance — check your key above, then contact whoever issued it for a top-up. Also fine to know: a base URL ending in /v1 works even though Anthropic clients append their own /v1 — the doubled path is handled.