DEV Community

Sonam
Sonam

Posted on

Build an AI Event Microsite on Telnyx Edge Compute

Most event apps quietly become several apps.

You start with a landing page. Then someone asks for SMS updates. Then attendees want a way to ask questions. Sponsors want lead capture. Organizers want feedback. Someone wants a daily report.

Suddenly your "event site" is a static frontend, a webhook server, a chatbot, a voice app, a database, an email job, and a few scripts holding it together.

The edge-event-microsite example compresses that into one deploy.

Code: https://github.com/team-telnyx/telnyx-code-examples/tree/main/edge-event-microsite

What it builds

This is a TypeScript app running on Telnyx Edge Compute.

It includes:

  • a server-rendered event microsite
  • event schedule and sponsor data stored in Telnyx KV
  • an SMS/WhatsApp AI concierge
  • in-browser voice AI with WebRTC
  • lead qualification with Telnyx AI Inference
  • voice feedback transcription
  • sponsor reporting
  • webhook signature verification

The important part is that all of these surfaces read from the same source of truth: one Telnyx KV namespace.

The architecture

At a high level:

Attendee browser
  -> Edge function
  -> Telnyx KV

SMS / WhatsApp message
  -> Edge function
  -> Telnyx AI Inference
  -> Telnyx Messaging
  -> Telnyx KV

Browser voice agent
  -> WebRTC
  -> AI assistant tool callback
  -> Edge function
  -> Telnyx KV
Enter fullscreen mode Exit fullscreen mode

The site, the text concierge, and the voice assistant all use the same event data.

Change the schedule once, and every surface can answer from the same updated state.

No sync jobs. No duplicated config. No stale FAQ hiding in a separate bot.

Why Edge Compute fits

An event experience has a lot of small, latency-sensitive interactions:

  • render the schedule
  • answer "where is the keynote?"
  • capture a lead
  • route a hot lead to sales
  • transcribe attendee feedback
  • generate a sponsor report

None of those need a giant backend.

They need a fast place to run close to the communications layer.

That is the useful pattern in this sample: the Edge function becomes the backend for the website, the messaging workflows, and the voice assistant tool calls.

KV as the shared event brain

The sample uses Telnyx KV for event state:

  • event details
  • schedule items
  • sponsor information
  • attendees
  • leads
  • feedback
  • assistant configuration

The microsite reads from KV when rendering pages and APIs.

The SMS/WhatsApp concierge reads from KV when answering attendee questions.

The voice assistant can call back into the same Edge function to look up event information.

That shared state is what keeps the experience coherent.

AI concierge for SMS and WhatsApp

Attendees can ask questions like:

Where is the keynote?
What time is the workshop?
Who is sponsoring the afterparty?
Enter fullscreen mode Exit fullscreen mode

The Edge function uses Telnyx AI Inference to generate a useful answer from the event context, then sends the response back with Telnyx Messaging.

The same flow can also qualify leads.

If a message suggests strong buying intent, the app can extract company, role, and interest level, store that lead in KV, and route the hot lead to a sales rep.

Voice AI in the browser

The sample also includes an in-browser voice path.

Instead of forcing attendees to call a phone number, the microsite can connect them to a Telnyx AI Assistant through WebRTC using @telnyx/ai-agent-lib.

The assistant is provisioned server-side and can use a webhook tool that queries the same KV data as the site and SMS concierge.

That gives attendees three ways to interact with the same event backend:

  • browse the site
  • text the concierge
  • talk to the assistant

Feedback transcription and sponsor reports

After a session, attendees can leave spoken feedback.

The app transcribes the audio, summarizes it with AI, and stores the result.

From there, the organizer can generate a sponsor report with lead counts, feedback summaries, and engagement metrics.

This is where the example starts to feel less like a website and more like an event operating system.

Run it

Clone the repo:

git clone https://github.com/team-telnyx/telnyx-code-examples.git
cd telnyx-code-examples/edge-event-microsite
npm install
Enter fullscreen mode Exit fullscreen mode

Authenticate and create a KV namespace:

telnyx-edge auth api-key set <YOUR_TELNYX_API_KEY>
telnyx-edge storage kv create --name edge-event-microsite-data
Enter fullscreen mode Exit fullscreen mode

Add webhook verification:

PUBLIC_KEY=$(curl -s -H "Authorization: Bearer $TELNYX_API_KEY" \
  https://api.telnyx.com/v2/public_key | jq -r '.data.public')

telnyx-edge secrets add TELNYX_PUBLIC_KEY "$PUBLIC_KEY"
Enter fullscreen mode Exit fullscreen mode

Deploy:

telnyx-edge ship
Enter fullscreen mode Exit fullscreen mode

The function deploys to a *.telnyxcompute.com URL.

Production notes

Before using this for a real event, I would add:

  • authentication for organizer/admin routes
  • stricter permissions around lead exports
  • rate limits for public messaging and voice flows
  • moderation for attendee-submitted feedback
  • consent language for transcription
  • retries for sponsor report delivery
  • analytics around which channel attendees prefer

The sample is intentionally compact, but the architecture is the interesting part.

One Edge function can serve the event website, answer attendee questions, power a voice assistant, capture feedback, and report back to sponsors.

Resources:

Top comments (0)