OneTag AI

Free LLM API: one HTTP POST, no credit card

Is there a free LLM API I can call from my code without a credit card?

Yes. OneTag AI is a free LLM API: one email address gets a key with 1,000,000 tokens a month and no credit card, and a call is a single HTTP POST with a prompt. There is no SDK to install and no model to pick.

The allowance starts again on the 1st of each month (UTC). One fast model answers every call. If you need to choose the model or keep the OpenAI request format, the comparison below lists the free LLM APIs that offer that, with their limits.

Steps

1. Get a free key

   Enter your email at onetagai.com. The key (it starts with ot_) arrives by email with a dashboard link. One key per email, no card.

2. Send a prompt

   One POST with the key, a prompt and an optional input:

   curl https://onetagai.com/v1/ask \
     -H "Content-Type: application/json" \
     -d '{"key": "YOUR_ONETAG_KEY", "prompt": "Summarize in one sentence", "input": "Paste any text here."}'

   The answer comes back as JSON with the tokens it used: {"text": "...", "usage": {"input": 61, "output": 14}}.

3. Call it from Python or Node

   Python needs only the standard library:

   import json, urllib.request
   
   def ask(prompt, input=None, **options):
       body = {"key": "YOUR_ONETAG_KEY", "prompt": prompt, "input": input, **options}
       req = urllib.request.Request(
           "https://onetagai.com/v1/ask",
           data=json.dumps(body).encode(),
           headers={"Content-Type": "application/json"},
       )
       with urllib.request.urlopen(req) as res:
           return json.load(res)["text"]
   
   print(ask("Classify the sentiment as positive, negative or neutral. One word.",
             "The checkout was fast and the box arrived a day early."))

   Node 18 or later has fetch built in; save this as a .mjs file. With json: true the model answers with a JSON object, returned as a string in text:

   const res = await fetch("https://onetagai.com/v1/ask", {
     method: "POST",
     headers: { "Content-Type": "application/json" },
     body: JSON.stringify({
       key: "YOUR_ONETAG_KEY",
       prompt: "Extract {name, price, currency}",
       input: "Aeropress Go, 39.95 USD, ships in two days.",
       json: true,
     }),
   });
   const { text, usage } = await res.json();
   console.log(JSON.parse(text), usage);

Request options

Streaming sends one data: line per piece of text, then a last event with done and the usage:

curl -N https://onetagai.com/v1/ask \
  -H "Content-Type: application/json" \
  -d '{"key": "YOUR_ONETAG_KEY", "prompt": "Write a haiku about static sites", "stream": true}'

data: {"delta": "Static"}

data: {"delta": " pages"}

data: {"delta": " load"}

...

data: {"done": true, "usage": {"input": 58, "output": 17}}

For a chat, send the earlier turns in history and the new message as the prompt:

curl https://onetagai.com/v1/ask \
  -H "Content-Type: application/json" \
  -d '{"key": "YOUR_ONETAG_KEY", "system": "Answer in one short sentence.", "prompt": "And in Fahrenheit?", "history": [{"role": "user", "content": "At what temperature does water boil at sea level, in Celsius?"}, {"role": "assistant", "content": "100 degrees Celsius."}]}'

Errors come back as JSON with an HTTP status, for example 401 with {"error": {"code": "bad_key", "message": "Unknown or disabled key."}}. The codes to handle are rate_limited and quota_exceeded (both 429), too_large (413) and upstream (502). The full reference is in the HTTP API docs.

Limits of the free key

From server code, the limit you meet first is the 20 requests a minute per IP address: a script on one machine cannot go faster than that, whatever the key allows.

Free LLM APIs compared

The table looks at each API the way code that calls it does: the request format, whether you choose the model, and what the free tier allows.

Other providers' cells were read on their own pages on 2026-11-10; follow the links for the current numbers. Free tiers change often.

Checked the same day and no longer free: GitHub Models (retired on July 30, 2026); Mistral (the free mode no longer includes API access; API keys need billing activated); Hugging Face Inference Providers (free accounts get no monthly credits); Cerebras (a $5 trial credit that expires after 30 days and needs a verified payment method).

Which free LLM API should you use?

• You want a specific open model, or to compare several: Groq or OpenRouter free models. Both accept the OpenAI request format, so code written for the OpenAI SDK works after changing the base URL and the key.
• You want Gemini models: Google's free tier. Check your limits in AI Studio first, because Google does not publish fixed numbers.
• You already deploy on Cloudflare: Workers AI.
• You want the least setup, or the call will run in a browser: OneTag AI. One POST, no model name, and the same key is safe in front-end code. See Free AI API for websites.

What OneTag AI does not do: there is no model parameter, no OpenAI-compatible route, no embeddings and no function calling. It is one endpoint that turns a prompt and an input into text or JSON.

The same key works in a web page

LLM keys are normally secrets that must stay on a server. A OneTag AI key is public by design: it can do one thing and can only spend its own token budget, so a static page can call the model with no backend. Add one script tag and call onetag.ask(prompt, input):

<script src="https://onetagai.com/v1.js" data-key="YOUR_ONETAG_KEY"></script>
<script type="module">
  const answer = await onetag.ask("Explain this in plain words", document.querySelector("article"));
</script>

Why that is safe, and what a stranger can do with a copied key: Is it safe to put an AI API key in a static HTML page?

Questions

Related guides

• Free AI API for websites: no backend, no credit card
• Free chatbot API for websites: chat history, streaming, no backend
• Free translation API for websites: no backend, no credit card
• Free OCR API: image to text from a web page, no backend
• Is it safe to put an AI API key in a static HTML page?

Updated 2026-11-11. Something wrong or missing? Email support@odysseyx.app.