The AskAI.free API

Quickstart

The endpoint speaks the same shape as the OpenAI chat completions API, so most existing clients work by changing two lines: the base URL and the key.

curl https://askai.free/v1/chat/completions \
  -H "Authorization: Bearer sk_askai_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-opus-5",
    "messages": [
      {"role": "system", "content": "You are a concise assistant."},
      {"role": "user", "content": "Summarise this quarter in one sentence."}
    ],
    "max_tokens": 500
  }'

# pip install openai
from openai import OpenAI

client = OpenAI(
    api_key="sk_askai_YOUR_KEY",
    base_url="https://askai.free/v1",
)

resp = client.chat.completions.create(
    model="claude-opus-5",
    messages=[{"role": "user", "content": "Summarise this quarter in one sentence."}],
    max_tokens=500,
)
print(resp.choices[0].message.content)

const res = await fetch("https://askai.free/v1/chat/completions", {
  method: "POST",
  headers: {
    "Authorization": "Bearer sk_askai_YOUR_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "chatgpt-5-6-sol",
    messages: [{ role: "user", content: "Write a haiku about deploys." }],
  }),
});
const data = await res.json();
console.log(data.choices[0].message.content);

What comes back

{
  "id": "chatcmpl-9f2c1ab4e7d3",
  "object": "chat.completion",
  "created": 1755820800,
  "model": "claude-opus-5",
  "choices": [
    { "index": 0, "message": { "role": "assistant", "content": "..." }, "finish_reason": "stop" }
  ],
  "usage": {
    "prompt_tokens": 42,
    "completion_tokens": 118,
    "total_tokens": 160,
    "askai_tokens_charged": 6400,
    "askai_tokens_remaining": 24993600
  }
}

askai_tokens_charged is what came off your monthly allowance: the provider's token count multiplied by that model's rate (see the table below). askai_tokens_remaining is what is left this cycle.

Models

Pass the id as model. The multiplier is how many allowance tokens one provider token costs - the same rate the website charges, so a chat on the site and the same call through the API cost exactly the same.

GET /v1/models returns the same list live, with an available flag per model for your plan.

Endpoints

Limits

• Rate: 60 per minute; 2000 per hour, counted per key.
• Input: up to 20,000 tokens per request on Pro, 80,000 on Max.
• Output: up to 4,000 tokens on Pro, 8,000 on Max. max_tokens is clamped to that ceiling rather than rejected.
• Streaming is not supported yet. Send "stream": false or omit it.

Running out of tokens

• Calls come out of the same monthly allowance as the website. There is no separate API balance, and nothing can be topped up mid-cycle.
• If your remaining balance cannot pay for the answer you asked for, max_tokens is quietly reduced to what it can cover, so a request never spends more than you have.
• At zero tokens, /v1/chat/completions returns 429 insufficient_tokens and no model is called. The allowance refills when your plan renews.
• /v1/usage and /v1/models keep answering at zero - they cost nothing, and your code needs them to see why it was cut off.

Errors

Errors use the standard shape, so existing client libraries surface them normally.

{ "error": { "message": "...", "type": "permission_error", "code": "trial" } }

Keeping your key safe

• The key is shown once, at creation. We store only a SHA-256 hash, so nobody - including us - can read it back.
• Keep it server-side. A key in browser JavaScript or a mobile app is a key anyone can copy and spend your allowance with.
• Lost it, or think it leaked? Revoke it on your account page and create a new one. Revocation takes effect on the next request.
• You can hold up to 5 active keys, so you can rotate without downtime.