Bolt.new AI and Webex by Cisco integration: Step-by-Step Guide 2025

To integrate Bolt.new with Webex by Cisco, you don’t integrate “Bolt itself.” Instead, inside a Bolt.new workspace you build an app (backend + frontend + AI agent flows) that talks to Webex using Webex’s real REST APIs and its OAuth or Bot Token authentication. The integration works exactly like any other external API call: you store credentials in Bolt.new environment variables, make HTTPS requests from your server code, and optionally expose public routes for Webex webhooks. Nothing automatic, nothing hidden — you wire it yourself using standard Webex API patterns.

 

What You Actually Do

 

You create a small backend inside Bolt.new (Node.js or Python). That backend will:

• Authenticate to Webex using Bot Token (simplest) or OAuth 2.0 (if you need user-level permissions).
• Call Webex REST APIs to send messages, read messages, or manage rooms.
• Expose a public webhook endpoint (Bolt gives your project a temporary public URL) to receive messages and events from Webex.
• Optionally call the Bolt AI/LLM runtime to generate responses to Webex messages.

 

How the whole flow works

 

Here’s the mental model that will never mislead you:

• Your Bolt.new backend acts like any normal server.
• Webex exposes public REST endpoints such as https://webexapis.com/v1/messages.
• You get a Bot Access Token from Webex Developer Portal and save it in your Bolt.new environment variables.
• Your code makes authenticated HTTP requests using that token.
• If you want Webex to call you, you create a Webex Webhook pointing to Bolt’s public endpoint.

 

Step‑by‑step practical instructions

 

This is the simplest and most reliable path for a first working integration.

• Create a Webex Bot in the Webex Developer Portal. You’ll receive a Bot Access Token.
• In your Bolt.new workspace, open the Environment Variables panel and add: WEBEX_BOT_TOKEN = your-token-here
• Inside Bolt.new, scaffold a small Node.js server (Express works well).
• Use Bolt’s built‑in public URL for webhook callbacks.

 

Minimal working Node.js example (Bolt.new server side)

 

// server.js
// A minimal Webex integration inside Bolt.new
import express from "express";
import fetch from "node-fetch";

const app = express();
app.use(express.json()); // Required for Webex webhook payloads

const WEBEX_TOKEN = process.env.WEBEX_BOT_TOKEN;

// Send a text message to a Webex room
async function sendMessageToWebex(roomId, text) {
  const response = await fetch("https://webexapis.com/v1/messages", {
    method: "POST",
    headers: {
      Authorization: `Bearer ${WEBEX_TOKEN}`,
      "Content-Type": "application/json"
    },
    body: JSON.stringify({
      roomId: roomId,
      markdown: text
    })
  });

  if (!response.ok) {
    console.error("Webex error:", await response.text());
  }
}

// Webhook endpoint for Webex to send events to
app.post("/webex/webhook", async (req, res) => {
  const data = req.body;

  // Webex webhook handshake
  if (data.resource === "messages" && data.event === "created") {
    const messageId = data.data.id;

    const msgRes = await fetch(`https://webexapis.com/v1/messages/${messageId}`, {
      headers: { Authorization: `Bearer ${WEBEX_TOKEN}` }
    });
    const message = await msgRes.json();

    // Ignore messages from the bot itself
    if (message.personEmail && !message.personEmail.includes("webex.bot")) {
      // Respond with a simple message
      await sendMessageToWebex(message.roomId, `You said: ${message.text}`);
    }
  }

  res.sendStatus(200);
});

// Health check
app.get("/", (req, res) => {
  res.send("Webex bot running inside Bolt.new");
});

app.listen(3000, () => console.log("Server started on port 3000"));

 

Registering the webhook in Webex (required)

 

Once your Bolt.new app is running, you get a public URL like:

https://your-bolt-project-id.bolt.live/webex/webhook

Go to Webex Developer Portal → Webhooks → Create Webhook. Use:

• Target URL: your Bolt public URL
• Resource: messages
• Event: created
• Include: all

Webex will now send message events to your Bolt app.

 

Connecting Bolt AI to Webex

 

If you want AI‑generated replies:

• Inside the /webex/webhook handler, make a call to Bolt’s LLM function.
• Generate text and send it back using sendMessageToWebex.

Example snippet inside that same route:

// Generate an AI reply inside Bolt
const aiReply = await bolt.ai.generate({
  prompt: `User said: "${message.text}". Generate a helpful answer.`
});

await sendMessageToWebex(message.roomId, aiReply);

 

Important operational notes

 

• Webex Bot Tokens never use OAuth redirect URLs. They’re static tokens.
• OAuth 2.0 is only required if you need user-level scopes (read/write all rooms).
• Bolt.new public URLs are temporary for development. For production you deploy to a stable hosting provider and update the Webex webhook.
• Webex requires HTTPS. Bolt.new public URLs are HTTPS by default, so you’re fine.

 

That is the real, working, correct way to integrate Bolt.new with Webex: you build a server inside Bolt, store your Webex token as an environment variable, call Webex APIs over HTTPS, and expose webhook endpoints that Webex can reach. Nothing hidden, everything explicit and standard.