How to Use Ox Alpha for Free (Browser Chat and API)

Ox Alpha is listed on OpenRouter at $0 per token during its preview, which is expected to run until around August 27, 2026. You do not need to pay anything, but you do need a free OpenRouter account and an API key. Here is the fastest path — in the browser, then via the API.

Option 1: chat in your browser (2 minutes)

1. Create a free account at openrouter.ai.
2. Open Settings → Keys and create a new API key. It starts with sk-or-. Copy it once — OpenRouter does not show it again.
3. Go to the Ox Alpha chat on this site, click API key, paste the key and save.
4. Start chatting. Responses stream in real time.

How the chat handles your key: it is stored only in your browser’s local storage and sent directly from your browser to OpenRouter. This website never receives your key, your prompts, or the model’s replies. Conversations are kept in memory and vanish when you leave the page or click New chat. You can remove the key at any time from the same panel.

Option 2: call the API directly

Ox Alpha speaks the standard OpenAI-compatible chat completions format through OpenRouter. Use the model ID stealth/ox-alpha:

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stealth/ox-alpha",
    "messages": [{"role": "user", "content": "Summarize this repository structure: ..."}]
  }'

Any OpenAI-compatible SDK works by pointing the base URL at https://openrouter.ai/api/v1. Tool calling, structured JSON output, and streaming are supported per the listing. Image and video inputs go in as standard multimodal content parts; audio is rejected.

Option 3: other gateways

The model is also available through OpenCode Zen, AI/ML API, AIHubMix, ModelsLab and Routin. Pricing and terms on those platforms are set by each platform — check them before use.

Tips for getting the most out of it

• Use the context. With 1,048,576 tokens available, paste whole files or folders rather than excerpts.
• Ask for long outputs. The 131,072-token output limit allows complete refactors or documents in one reply.
• Expect rate limits. Free stealth previews are shared by many users; a 429 response means wait and retry.
• Keep it non-sensitive. Prompts may be retained by the anonymous provider under OpenRouter’s Stealth Program terms (the listing says not for training). No personal data, secrets, or confidential material.
• Watch the clock. When the preview ends the model may go paid, be renamed, or disappear. Our tracker and the banner on the chat page update automatically.

FAQ

Do I need to add credits to OpenRouter?

Not for Ox Alpha while it is listed at $0 per token. If the model moves to paid pricing, requests will draw on your OpenRouter balance.

Is there a usage limit?

OpenRouter applies rate limits to free models. They are not published per model; if you hit a 429 error, back off and retry.

Can I use it in my app?

Yes, via the API, but remember it is a pre-release preview with no guarantee of availability. Build in a fallback model.

What happens after the preview?

Based on every previous stealth model, it will be revealed under its official name and typically re-listed with paid pricing. See the reveal history.