# Modelane Quickstart

Modelane is a unified inference platform for global teams. Use one account, one billing layer, and one governed API surface across configured frontier model families. Modelane is operated by XNODEHUB ENTERPRISES PTE. LTD., Singapore.

TL;DR:

1. Sign in to the Modelane console.
2. Create an API key.
3. Use `https://api.modelane.ai/v1` for OpenAI-compatible clients.
4. Use `https://api.modelane.ai/v1/messages` for Anthropic Messages-compatible clients.
5. Use `https://api.modelane.ai/v1beta/models/{model}:generateContent` for Gemini-compatible clients.

Public evaluation links:

- Pricing: https://modelane.ai/pricing
- Status: https://modelane.ai/status
- Prompt Savings Report: https://modelane.ai/savings
- Affiliates: https://modelane.ai/affiliates
- LLM index: https://modelane.ai/llms.txt

## OpenAI-Compatible API

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.modelane.ai/v1",
    api_key="ml-your-key",
)

response = client.chat.completions.create(
    model="modelane-fast",
    messages=[{"role": "user", "content": "Hello, Modelane."}],
)

print(response.choices[0].message.content)
```

## Claude Code

macOS / Linux:

```bash
curl -fsSL https://modelane.ai/install/claude-code.sh | bash
claude-modelane "Summarize this repository"
```

Windows PowerShell:

```powershell
irm https://modelane.ai/install/claude-code.ps1 | iex
powershell -File $HOME\.modelane\bin\claude-modelane.ps1 "Summarize this repository"
```

The wrapper sets `ANTHROPIC_BASE_URL=https://api.modelane.ai` and sends the Modelane key as `ANTHROPIC_AUTH_TOKEN` only for that Claude Code process.

## Codex CLI

macOS / Linux:

```bash
curl -fsSL https://modelane.ai/install/codex.sh | bash
codex-modelane "Review the current diff"
```

Windows PowerShell:

```powershell
irm https://modelane.ai/install/codex.ps1 | iex
powershell -File $HOME\.modelane\bin\codex-modelane.ps1 "Review the current diff"
```

The installer creates `~/.codex/modelane.config.toml` and runs Codex with `--profile modelane`, leaving the user's default Codex provider untouched.

## Gemini CLI

macOS / Linux:

```bash
curl -fsSL https://modelane.ai/install/gemini-cli.sh | bash
gemini-modelane -p "Explain this project"
```

Windows PowerShell:

```powershell
irm https://modelane.ai/install/gemini-cli.ps1 | iex
powershell -File $HOME\.modelane\bin\gemini-modelane.ps1 -p "Explain this project"
```

The wrapper sets `GEMINI_API_KEY`, `GOOGLE_GEMINI_BASE_URL=https://api.modelane.ai`, and `GEMINI_MODEL` only for that Gemini CLI process.

## Direct REST Examples

Claude Messages-compatible:

```bash
curl https://api.modelane.ai/v1/messages \
  -H "x-api-key: ml-your-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "max_tokens": 256,
    "messages": [{"role": "user", "content": "Hello, Modelane."}]
  }'
```

Gemini-compatible:

```bash
curl "https://api.modelane.ai/v1beta/models/gemini-2.5-flash:generateContent?key=ml-your-key" \
  -H "content-type: application/json" \
  -d '{
    "contents": [{
      "parts": [{"text": "Hello, Modelane."}]
    }]
  }'
```

## Notes

- Model availability depends on account configuration and routing policy.
- Billing is prepaid balance based. Review usage and pricing in the console.
- Keep API keys private. Do not paste keys into public repositories or issue trackers.
- Modelane accounts are governed by the Modelane Terms, Privacy Policy, and Acceptable Use Policy.
- Referral links use the `?aff=CODE` sign-up parameter and current rewards are wallet-balance oriented.
- Use the Prompt Savings Report page for a first-pass cost comparison, then confirm exact rates on the live pricing page.