Quickstart · developers and agents

From zero to inference, without leaving this page.

No signup, no card, no waitlist. Press Run: a free key is minted for this browser, a real completion runs on the network, and every snippet below is rewritten with your key.

The endpoint is OpenAI-compatible: anything written against the OpenAI SDK works by changing one base URL.

myai quickstart · liveno key yet

# press Run and watch the quickstart happen for real

Mints a free key for this browser, then a real completion. ~2 s.

1Get a key

request
curl -X POST https://api.myaitoken.io/v1/keys/auto-issue \
  -H 'Content-Type: application/json' \
  -d '{"machine_id":"my-laptop"}'
response
{
  "api_key": "myai-sk-...",
  "plan": "free",
  "free_credit_usd": 5.0,
  "rate_limit_rpm": 10
}

Anonymous keys work immediately. Add "email" to the body (or call /v1/keys/link-email later) when you want the key to be billable past the free credit.

2Make your first call

request
curl -X POST https://api.myaitoken.io/v1/chat/completions \
  -H "Authorization: Bearer myai-sk-..." \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "qwen3:8b",
    "messages": [{"role":"user","content":"Say hello"}]
  }'

The response includes chosen_agent_name, the machine that ran your job. Every completion is real work on someone's GPU.

3Or point your existing code at us

python
from openai import OpenAI

client = OpenAI(
    api_key="myai-sk-...",
    base_url="https://api.myaitoken.io/v1",   # the only line that changes
)

print(client.chat.completions.create(
    model="qwen3:8b",
    messages=[{"role": "user", "content": "Say hello"}],
).choices[0].message.content)
node
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "myai-sk-...",
  baseURL: "https://api.myaitoken.io/v1",
});

What to know before you build

  • Free keys have a published limit of 10 requests/minute; API-level enforcement is not currently enabled and limits may change. Link an email for higher limits and billing.
  • See available models or call GET /v1/models; the catalog reflects what providers are actually serving.
  • Warm requests typically return in about a second; a cold provider spin-up can take several seconds. Honest numbers in the OpenAI comparison.
  • AI agents get a 15% discount versus human buyers. Details in the docs.

When you outgrow the free credit

Attach an email to the key you already have. We email that address a confirmation link; opening it links the key and issues it a fresh secret (update your config), and purchased credit applies to it from then on:

make an existing key billable
curl -X POST https://api.myaitoken.io/v1/keys/link-email \
  -H 'Content-Type: application/json' \
  -d '{"api_key":"myai-sk-...","email":"[email protected]"}'

Then buy a job pack from pricing: pay-per-job from $0.02 (small models), $0.08 (medium), $0.25 (large). No subscription, no setup fee.

Read the docsSee pricingOr provide compute