04 / EMBED
One snippet. Your website.
Choose your widget design and paste a lightweight script into WordPress, React, or your website layout. No jQuery required.
The intelligence layer for your product
Bring intelligent conversations into your applications with a focused LLM chat API. Or add a website widget that connects AI answers with real human support.
For developers, startups and modern software teams. Text chat only.
ONE API. MORE POSSIBILITIES.
The platform
One text chat platform. Two ways to build: use the API directly, or bring AI and human support to your website.
04 / EMBED
Choose your widget design and paste a lightweight script into WordPress, React, or your website layout. No jQuery required.
01 / INTEGRATE
Call the text chat API from your backend. Keep your own interface, language, and application logic.
02 / CONVERSE
Embed a branded widget with AI answers, project knowledge, and a human handoff to your Pixizen Inbox.
03 / MEASURE
Follow input, cached input, output, and spend. Set key budgets and receive realtime usage events.
10 widget design presets
Knowledge file uploads
AI ↔ human takeover
Separate API wallet
Focused on text conversations. Image, video and audio generation are separate Pixizen products.
Interactive preview
Try a local example of the experience. The real widget connects your visitors to project knowledge, AI responses, and your team's inbox.
LOCAL SIMULATION — NO BACKEND CONNECTION
How can I add AI chat to my app?
Choose the server API for your own interface, or install the website widget for AI plus human support. Create a project, keep credentials server-side, and follow the quickstart. This is a simulated example.
Simulation only. No API request, payment, or personal data submission.
A small snippet. A complete conversation.
Your website, your branding, your customer conversations. Install once in your shared layout and carry your support experience across every page.
YOUR PUBLIC EMBED · EXAMPLE
<script src="https://YOUR_BACKEND/api/chat-widget/v1.js" data-widget-id="YOUR_PUBLIC_WIDGET_ID" async></script>
AI + your team, together
Try the conversation
Type a message and explore the widget experience. This guided demo runs locally: no real agent, API charge, or external message is involved.
Local preview · not a live support session
Hello! I’m the demo assistant. Ask about widgets, knowledge, or human support.
How it works
Start with an account, choose the integration you need, and keep everything under your Pixizen workspace.
Open the API workspace inside your Pixizen account.
Create a server API key or configure a website widget.
Keep secret keys on your server. Embed only the public widget ID.
Send a text request or let a visitor message your widget.
Receive streaming replies and monitor usage in your dashboard.
Developer experience
Use a familiar HTTP request from your server. Stream replies with SSE, or receive a complete JSON response. Your application talks to Pixizen; provider credentials stay on our backend.
# Run on your server. Use your deployed Pixizen backend URL.
curl "$PIXIZEN_BASE_URL/chat/completions" \
-H "Authorization: Bearer $PIXIZEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "pixil-xd-1.3",
"messages": [{"role":"user","content":"Hello!"}],
"max_tokens": 300,
"stream": false
}'// Server-side JavaScript. Never expose this key in a browser.
const response = await fetch(
process.env.PIXIZEN_BASE_URL + '/chat/completions', {
method: 'POST',
headers: {
Authorization: 'Bearer ' + process.env.PIXIZEN_API_KEY,
'Content-Type': 'application/json'
},
body: JSON.stringify({
model: 'pixil-xd-1.3',
messages: [{ role: 'user', content: 'Hello!' }],
max_tokens: 300, stream: false
})
});
if (!response.ok) throw new Error('Chat request failed');
const data = await response.json();
console.log(data.choices[0].message.content);# Server-side Python. Store your key in environment variables.
import os, json, urllib.request
request = urllib.request.Request(
os.environ['PIXIZEN_BASE_URL'] + '/chat/completions',
data=json.dumps({
'model': 'pixil-xd-1.3',
'messages': [{'role': 'user', 'content': 'Hello!'}],
'max_tokens': 300, 'stream': False
}).encode(),
headers={
'Authorization': 'Bearer ' + os.environ['PIXIZEN_API_KEY'],
'Content-Type': 'application/json'
}, method='POST')
with urllib.request.urlopen(request, timeout=100) as response:
print(json.load(response)['choices'][0]['message']['content'])What you can build
Use the hosted widget for website support, or build your own text experiences on the API.
Add a floating AI support widget to your website, WordPress store, or web app.
Use the text API from your backend to build assistance into your own app.
Ground answers in uploaded knowledge, then let a person take over in Inbox.
Build conversational product features with project-scoped API keys.
Create text assistants for your team with explicit budgets and access limits.
Stream language-model output into your own business logic and interfaces.
Pricing & testing
See exactly what goes into a reply. Uncached input, reported cached input, and generated output are measured separately.
These are the current test configuration rates, not a promise of future live pricing. Your account shows its applicable rate version. API balance is separate from image and video credits.
$0 subscription · $10 test credit · 30 days.
Granted by a Superadmin. No real payment is taken in test mode.
| Token type | Rate / 1M |
|---|---|
| Uncached input | $1.25 |
| Cached input | $0.15 |
| Output | $12.50 |
Cached input is included in total input; it is never counted twice. Missing cache metering is marked unavailable. Test credit cannot spend your paid balance.
Developer resources
Go from your first request to streaming, widget installation, token accounting, and production configuration.
Create a project, generate a key, and send your first request.
READ GUIDEUse project-scoped keys and keep credentials on your server.
READ GUIDEWork with messages, JSON responses and SSE streaming.
READ GUIDEUnderstand retry rules, token reservations and spending limits.
READ GUIDEFAQ
A few helpful answers about integrating the text API, installing your widget, and moving between AI and human support.
A text-only LLM chat platform with two integration paths: a server API for your own applications, and an embeddable website widget with AI and human support. All requests go through the Pixizen backend.
Yes. Create a project and a secret API key, then call the chat endpoint from your backend. A website widget is optional, and API-only access does not require widget domain verification.
Create a project, verify your exact HTTPS hostname with a DNS TXT record, choose one of ten presets, and copy the small script into your shared website layout or WordPress footer. No jQuery is required.
Yes. A visitor can request human support. An authorized operator takes over from the Pixizen Inbox, which pauses AI delivery. Returning the conversation to AI is an explicit action.
Uncached input, reported cached input, and generated output have separate rates. Cached tokens are part of total input, not extra tokens. If the provider does not report cache usage, the dashboard marks it unavailable and no cached discount is assumed.
No. A Superadmin grants the free test subscription. The current test package includes $10 of separate test credit for 30 days at a $0 subscription price. Test mode does not take real payments or spend your paid balance.
Yes. Upload TXT or Markdown, or extract text from a PDF in the customer panel. The widget retrieves relevant project knowledge. Scanned PDFs need OCR first; API-only requests currently use the messages you supply.
No. This product and its keys are limited to text chat. Pixizen's creative products and their credits are separate. Native mobile apps need a WebView or a backend integration; a native mobile SDK is not included.
The developer guide covers getting started, Bearer authentication, cURL/JavaScript/Python examples, SSE streaming, notification replay, error handling, token accounting, budgets, and production configuration.
Build with Pixizen
Build your own chat experience with the API, or connect your website visitors to AI and your team with a single widget.
One Pixizen account. Separate Chat API billing. Your choice of integration.