Works with Twilio SMS & WhatsApp

AI Chatbot for Twilio: turn a phone number into a support line that never sleeps

Your Twilio number can already receive texts. What it cannot do is answer them. Connect it to Asyntai and every SMS gets a real reply — drawn from your own website content, in the customer's language, at 3am on a Sunday. The whole thing is a Twilio Studio flow with one HTTP widget. No server, no middleware.

Get Started Free See the setup

See what the chatbot would answer

Enter your website and watch the AI answer questions from your own content — the same answers it would text back.

Why an SMS chatbot beats another inbox

Text messages get opened. That is the whole problem with them — a Twilio number attached to a small team turns into a queue of "is this still available?" and "what time do you close?" that someone has to work through by hand, usually the owner, usually after hours.

The questions are rarely hard. They are the same twenty questions your website already answers, arriving one at a time through a channel with no FAQ page attached. A chatbot on the other end of the number changes the arithmetic: the routine texts get answered in seconds, and the ones that genuinely need a person are the only ones that reach you.

Same number, same customers — the difference is that nobody has to be holding the phone.

How the Twilio chatbot works

Twilio Studio is the visual flow builder in your Twilio console. It can call an external API and use the reply, which is all this integration needs.

1. Trigger
A customer texts your Twilio number
2. HTTP request
Studio POSTs the message to the Asyntai API
3. Answer
Asyntai replies using your site content
4. Send
Studio texts the answer back

The three steps in Twilio Studio

Get your API key. In Asyntai, open Settings → API and generate a key. API access is included on paid plans.
Add a Make HTTP Request widget. Connect it to the Incoming Message trigger and name it asyntai. Set method POST, content type Application/JSON, and paste the URL below — your key goes in the query string, because Studio's HTTP widget does not send custom headers.
Reply with the answer. Add a Send Message widget with body {{widgets.asyntai.parsed.response}}. Publish the flow and point your phone number at it in the Twilio console.
Request URL
https://asyntai.com/api/v1/chat/?api_key=YOUR_API_KEY
Request body
// Body = what the customer texted // From = their number, so the chatbot remembers the thread { "message": "{{trigger.message.Body}}", "session_id": "{{trigger.message.From}}" }
What the Asyntai API returns
{ "success": true, "response": "We're open until 6pm on Saturdays.", "session_id": "+15551234567" }

The response string is what your customer receives. If you would rather keep the key out of the URL, swap the HTTP widget for a Twilio Function and send the key as an X-API-Key header instead — the API accepts both. Full endpoint reference lives in the API documentation.

The other direction: text yourself when it matters

Traffic does not only flow toward the chatbot. Asyntai webhooks fire an escalation.requested event the moment a visitor asks for a human, plus events for new conversations and agent takeovers. Point that webhook at a Twilio Studio flow's REST trigger and the right person gets an SMS with the question, while the visitor is still on the page.

That is the pattern worth setting up even if you never run the SMS bot: your website chat stays automated, and the handful of conversations that need you actually reach your pocket.

What your Twilio chatbot can do

Answers from your own content

The chatbot reads your website, plus any documents or spreadsheets you upload, and answers from that — not from generic AI knowledge.

Remembers the thread

Pass the sender's number as session_id and follow-up texts are understood in context, the way a real conversation works.

SMS and WhatsApp, one flow

Twilio Studio treats both channels the same, so a single flow covers texting customers and WhatsApp customers alike.

Replies in their language

The chatbot detects the language of the incoming message and answers in it — 30+ languages, nothing to configure.

Twilio chatbot questions

How do I connect an AI chatbot to Twilio?

Build a Studio Flow on the Incoming Message trigger, add a Make HTTP Request widget POSTing to https://asyntai.com/api/v1/chat/?api_key=YOUR_API_KEY, then send the returned response back with a Send Message widget. Studio's HTTP widget cannot send custom headers, which is why the key sits in the query string. No server or middleware required.

Does it remember the conversation?

Yes, when you pass the sender's number as session_id. Asyntai keeps history per session, so "and what about delivery?" is understood as a follow-up rather than a fresh question.

Does this work for WhatsApp as well as SMS?

Yes — Twilio Studio handles both with the same flow. If WhatsApp is your main channel, we also have a direct WhatsApp integration that skips Twilio entirely.

What does it cost?

The Asyntai API is included on paid plans (Starter, Standard, Pro) — see pricing. The free plan's 100 monthly messages do not include API access. Twilio bills you separately for the number and per-message SMS fees.

Can Twilio text me when a visitor needs a human?

Yes. The escalation.requested webhook fires when someone asks for a person; route it to a Studio REST trigger or your own endpoint that calls Twilio, and the on-call agent gets the question by SMS.

Other channels

Asyntai answers on the web, in messaging apps, and anywhere the API reaches.

Give your Twilio number something to say

Connect Asyntai to Twilio and every text gets answered from your own content — day, night, and weekends included.

Get Started Free