Embedded flowise is not working

[TVNCH]

I embedded the agent flow to a website. first the chatbot takes too much time to reply, second even if i upload a document it doesnt read. I have flowise free account. any suggestions?

[xXMrNidaXx]

Two issues here — slow response + document not reading:

1. Slow response fixes:

• Cold start on free tier — First request after idle is slow. Add a warmup ping:

  // Ping on page load to warm up
  fetch("https://your-flowise.cloud/api/v1/prediction/xxx", {
    method: "POST",
    body: JSON.stringify({ question: "ping" })
  });

• Model latency — If using GPT-4, switch to GPT-3.5-turbo for faster responses

• Vector store query — Large knowledge bases take longer. Reduce top-k results.

2. Document upload not working:

• Check the flow — Ensure Document Loader → Text Splitter → Vector Store chain is connected

• Embedding model — Verify your embedding API key is set

• File type — Some file types need specific loaders (PDF Loader, CSV Loader, etc.)

• Free tier limits — Check if you hit upload/storage limits

Debug steps:

1. Test the flow directly in Flowise UI (not embedded)
2. Upload a small test file (< 1 page PDF)
3. Check execution logs for errors

Embedding code check:

<script type="module">
  import Chatbot from "https://cdn.jsdelivr.net/npm/flowise-embed/dist/web.js";
  Chatbot.init({
    chatflowid: "your-id",
    apiHost: "https://your-flowise.cloud",
    // Ensure streaming is enabled
    streaming: true,
  });
</script>

We deploy embedded Flowise at Revolution AI — the warmup ping + streaming mode makes a big difference.