OneTag AI

Free chatbot API for websites: chat history, streaming, no backend

Is there a free chatbot API I can put on my website without a backend?

Yes. OneTag AI is a free chatbot API that a web page can call directly: each email address gets a key with 1,000,000 tokens a month and no credit card, and one call takes the new message, the earlier turns and a system prompt, and can stream the answer. The key is safe to leave in page source, so the chatbot needs no backend.

The allowance starts again on the 1st of each month (UTC). Below is a complete chat box in one HTML file, a one-line widget if you do not want to write code, what a conversation costs, and a comparison with the other free AI APIs.

Steps

1. Get a free key

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

2. Paste the chat box

   This is a working chatbot: a form, a log, and one call per message. system sets the bot's role, history carries the earlier turns, and stream shows the answer as it is written:

   <div id="log"></div>
   <form id="chat">
     <input id="q" placeholder="Ask something" autocomplete="off" required>
     <button>Send</button>
   </form>
   
   <script src="https://onetagai.com/v1.js" data-key="YOUR_ONETAG_KEY"></script>
   <script type="module">
     const log = document.getElementById("log");
     const q = document.getElementById("q");
     const history = [];
     const line = (who) => {
       const el = log.appendChild(document.createElement("p"));
       el.textContent = who + ": ";
       return el;
     };
   
     document.getElementById("chat").onsubmit = async (event) => {
       event.preventDefault();
       const question = q.value;
       q.value = "";
       line("You").textContent += question;
       const bot = line("Bot");
       try {
         const answer = await onetag.ask(question, null, {
           system: "You are the assistant of Acme Bikes, a bike shop. Answer in under 60 words.",
           history: history.slice(-20),
           stream: (delta) => { bot.textContent += delta; },
         });
         history.push({ role: "user", content: question }, { role: "assistant", content: answer });
       } catch (e) {
         bot.textContent += e.code === "quota_exceeded" ? "The chat is resting for now." : e.message;
       }
     };
   </script>

3. Make it yours

   Change the system line to describe your site and how the bot should answer. To let it answer from your own content, pass that content as the input: onetag.ask(question, document.querySelector("main"), { system, history }).

No code: a floating chat button

If you only want a chat button that answers questions about the page it is on, add this one tag instead. It adds a floating button and a chat panel, and needs no other code:

<script async src="https://onetagai.com/widget.js" data-key="YOUR_ONETAG_KEY"></script>

Options such as the title, greeting, colour and position are data- attributes; they are listed in the docs.

Call it over HTTP

For a bot that runs on a server, such as a Discord or Telegram bot, the same key works with one POST. 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."}]}'

Python and Node examples and the streaming format are on Free LLM API.

What a conversation costs

In our test on 2026-11-10, one turn with a system line, two earlier messages and a one-sentence answer used 124 tokens (108 in, 16 out). 1,000,000 tokens is about 8,000 turns of that size; longer answers and longer histories get fewer.

The API keeps no conversation state. The page sends the earlier turns again with every message, so each turn costs more than the one before. The example sends only the last 20 messages, which caps the cost of a long conversation.

What the free key includes

Free AI APIs for a chatbot, compared

Any chat model API can power a chatbot. The difference for a website is where the key lives: most free tiers give you a secret key that needs a server in front of it.

Other providers' figures were read on their own pages on 2026-11-10; follow the link in each row 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).

Limits to know before you rely on it

• The bot knows only what is in the system prompt, the history and the input you pass, and it will answer confidently about things you never told it. Put the facts it needs (opening hours, prices, policies) in the system prompt or the input. There is no training step and no file upload; to answer from a whole site, retrieve the relevant pages first, as in Add 'ask the docs' AI search.
• It does not remember visitors between page loads. Save history in localStorage if the conversation should survive a reload.
• Up to 100 history messages and 32,000 characters of text per call.
• One fast model answers; there is no model parameter and no function calling.
• Anyone can read the key in the page source and use it elsewhere. They can only spend the key's budget, and calls are limited to 60 a minute per key and 20 a minute per IP address. Pro keys can be locked to your domain.

Questions

Related guides

• AI chatbot in HTML, CSS and JavaScript: complete working code
• Chatbot HTML templates: 3 free chat UIs with working AI
• How to add an AI chatbot to an HTML website with one script tag
• Add 'ask the docs' AI search to a documentation site
• Free AI API for websites: no backend, no credit card
• 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.