Configuration

Create and copy your API key

Choose a plan on the pricing page, then open API Keys in your dashboard. Create an API key, select the AWSQ or CCMAX channel for Claude Code, and set its quota and expiration as needed. If you already have an API key, you can use it directly.

claude-opus-5-5 and claude-sonnet-5-5 are available on AWSQ and CCMAX. Requests can also use the aliases claude-opus-5.5 and claude-sonnet-5.5. Opus 5.5 base rates per 1M tokens are $4 input, $0.20 cache read, and $20 output; AWSQ consumes quota at 0.8× and CCMAX at 2×.

If the upstream actually uses Fast mode, Opus 5.5 base and cache rates double. Billing follows the response’s usage.speed. See Anthropic Fast mode pricing.

A generation request is allowed while the API key and its funding source have a positive remaining balance. Output and tool limits are independent of that balance. Each completed request is billed from actual usage, so concurrent requests can take the balance below zero.

For native Claude Messages, send the full conversation and inline file contents on each turn. A thread continuation returns HTTP 404 with thread_not_found, so clients can replay their local history. Shared upstream file and container IDs cannot be used.

Shared credentials support Anthropic-provided Skills. Private Skills, remote MCP servers and private tunnels need tenant-scoped authorization.

For predictable costs, set max_uses explicitly. Web Search accepts 1–20 calls. Advisor accepts at most two calls and 1,024–4,096 output tokens per call when those limits are specified. Advisor supports Web Search and client tools.

Standalone code execution costs about $0.00417 per activated container before the channel multiplier ($0.05/hour × 5 minutes). Multiple tool calls in one container incur that fee once. Including web_search_20260209, web_fetch_20260209 or a later version waives the container fee.

When declarations can be adjusted, the proxy defaults to one fetch with 4,096 text tokens per result, five web searches, and two Advisor calls with 4,096 output tokens each. PDFs are outside the fetch text cap. Requests with bound thinking, mixed tools or a pending tool call keep their declared limits. Mythos advisor models are unsupported.

Client tools on shared credentials must be directly callable. Programmatic calls that would pause a container for a client result are detected before generation. Pre-output refusals in the bio, frontier_llm and reasoning_extraction categories incur input charges. Fallback charges use each model’s actual billable usage.

Use the copy button:

In every example below, fill in YOUR_DASHBOARD_API_KEY with the key you copied before saving the configuration or running the commands.

Which Base URL should I use?

Claude Code appends /v1/messages to the base URL. The older /api/anthropic prefix is supported for existing integrations. Existing Base URLs ending in /api/codex or /api/gemini continue to work with the same keys.

Anthropic / Claude Code

For AWSQ and CCMAX keys, set ANTHROPIC_BASE_URL to https://api.aiinide.com. Claude Code and Anthropic SDKs append /v1/messages; do not add /v1 to this setting. OpenAI-compatible clients use https://api.aiinide.com/v1 instead.

Copy
1
curl --globoff 'https://api.aiinide.com/v1/messages' \
2
  -H "x-api-key: YOUR_DASHBOARD_API_KEY" \
3
  -H "anthropic-version: 2023-06-01" \
4
  -H "Content-Type: application/json" \
5
  -d '{"model":"claude-haiku-4-5-20251001","messages":[{"role":"user","content":"Hello"}],"max_tokens":256}'

Add "stream":true for streaming. GET /v1/models returns the Anthropic model fields and cursor pagination when the anthropic-version header is present; without it, the endpoint returns the OpenAI format. The same AWSQ or CCMAX key works with either protocol. With Codex or Gemini keys, omit anthropic-version on GET /v1/models. Including it returns 400.

POST /v1/messages/count_tokens retains the existing compatibility response (input_tokens: 1); it does not provide an accurate token count, call upstream, or consume quota.

OpenAI-compatible clients

AWSQ and CCMAX keys can also call Claude from OpenAI-compatible clients:

• Base URL: https://api.aiinide.com/v1
• API key: your dashboard AWSQ or CCMAX key
• Model: for example, claude-haiku-4-5-20251001; query GET /v1/models for your key's model list
• Endpoint: POST /v1/chat/completions or POST /v1/responses, with streaming or non-streaming responses

Copy
1
curl --globoff 'https://api.aiinide.com/v1/chat/completions' \
2
  -H "Authorization: Bearer YOUR_DASHBOARD_API_KEY" \
3
  -H "Content-Type: application/json" \
4
  -d '{"model":"claude-haiku-4-5-20251001","messages":[{"role":"user","content":"Hello"}],"max_tokens":256}'

For Responses, use input and max_output_tokens. Add "stream":true to stream. Client-executed function tools and multi-turn tool results are supported. Channel multipliers, membership discounts and quota limits still apply. Send the complete conversation each time; background generation, stored context, hosted tools, Fast mode and one-hour cache extensions are not supported here. Responses uses store=false. Use Anthropic Messages for Claude extensions such as Fast mode and hosted tools.

Configure required environment variables

AI in IDE needs four environment variables before the CLI or VS Code extension can talk to the proxy. Set them once, then restart your shell or editor so every new session picks them up.

macOS or Linux

Add the exports below to your shell profile (~/.zshrc for zsh, ~/.bashrc for bash):

Copy
1
export ANTHROPIC_BASE_URL='https://api.aiinide.com'
2
export ANTHROPIC_AUTH_TOKEN="YOUR_DASHBOARD_API_KEY"
3
export API_TIMEOUT_MS=600000
4
export CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1

After saving the file, reload it with source ~/.zshrc.

Prefer to append commands manually? Run them one at a time (swap the path if you use bash):

Copy
1
printf 'export ANTHROPIC_BASE_URL=%q\n' 'https://api.aiinide.com' >> ~/.zshrc
2
echo 'export ANTHROPIC_AUTH_TOKEN="YOUR_DASHBOARD_API_KEY"' >> ~/.zshrc
3
echo 'export API_TIMEOUT_MS=600000' >> ~/.zshrc
4
echo 'export CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1' >> ~/.zshrc
5
 
6
source ~/.zshrc

Windows

You can define the same variables through whichever flow you prefer:

Option 1 · GUI (recommended for permanent setup)

1. Press Win + R, type sysdm.cpl, press Enter.
2. Open the Advanced tab → Environment Variables…
3. Under User variables click New… and add each pair:
   • ANTHROPIC_BASE_URL → https://api.aiinide.com
   • ANTHROPIC_AUTH_TOKEN → YOUR_DASHBOARD_API_KEY
   • API_TIMEOUT_MS → 600000
   • CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC → 1
4. Close the dialogs and restart your terminal or IDE.

Option 2 · Command Prompt (persists)

Copy
1
setx ANTHROPIC_BASE_URL "https://api.aiinide.com"
2
setx ANTHROPIC_AUTH_TOKEN "YOUR_DASHBOARD_API_KEY"
3
setx API_TIMEOUT_MS "600000"
4
setx CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC "1"

Option 3 · PowerShell (persists)

Copy
1
[System.Environment]::SetEnvironmentVariable('ANTHROPIC_BASE_URL', 'https://api.aiinide.com', 'User')
2
[System.Environment]::SetEnvironmentVariable('ANTHROPIC_AUTH_TOKEN', 'YOUR_DASHBOARD_API_KEY', 'User')
3
[System.Environment]::SetEnvironmentVariable('API_TIMEOUT_MS', '600000', 'User')
4
[System.Environment]::SetEnvironmentVariable('CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC', '1', 'User')

Option 4 · Command Prompt (temporary—current window only)

Copy
1
set "ANTHROPIC_BASE_URL=https://api.aiinide.com"
2
set ANTHROPIC_AUTH_TOKEN=YOUR_DASHBOARD_API_KEY
3
set API_TIMEOUT_MS=600000
4
set CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1

Start the AI in IDE coding environment

Run the CLI inside any project where you want to use AI coding assistance:

Copy
1
cd your-project
2
claude

You should see a confirmation screen similar to:

Prefer the VS Code (or other IDE) extension? Install a compatible extension and open the command palette. If the extension prompts for configuration like this:

your environment variables were not detected.

Double‑check the shell setup, or hard-code the configuration for the extension:

• CLI settings file

  • macOS/Linux: ~/.claude/settings.json
  • Windows: C:/Users/<UserName>/.claude/settings.json

  Update it to include:

  Copy
  1
  {
  2
    "env": {
  3
      "ANTHROPIC_AUTH_TOKEN": "YOUR_DASHBOARD_API_KEY",
  4
      "ANTHROPIC_BASE_URL": "https://api.aiinide.com",
  5
      "API_TIMEOUT_MS": "600000",
  6
      "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1"
  7
    }
  8
  }

• VS Code workspace/user settings – add the snippet below to settings.json:

  Copy
  1
  "claudeCode.environmentVariables": [
  2
    { "name": "ANTHROPIC_BASE_URL", "value": "https://api.aiinide.com" },
  3
    { "name": "ANTHROPIC_AUTH_TOKEN", "value": "YOUR_DASHBOARD_API_KEY" }
  4
  ],

After editing either file, restart VS Code (or reload the window) so it picks up the new values.